(case study) what is margaret's current bmi

• Subscribe to journal Subscribe
• Get new issue alerts Get alerts

Secondary Logo

Journal logo.

Colleague's E-mail is Invalid

Your message has been successfully sent to your colleague.

Save my selection

Body Mass Index

Obesity, bmi, and health, a critical review.

Nuttall, Frank Q. MD, PhD

Frank Q. Nuttall, MD, PhD, is a full professor at the University of Minnesota, Minneapolis, and chief of the Endocrine, Metabolic and Nutrition Section at the Minneapolis VA Medical Center, Minnesota. His PhD degree is in biochemistry. He has more than 250 scientific publications in peer-reviewed journals, and he is the winner of numerous prestigious academic and scientific awards, including the 2014 Physician/Clinician Award of the American Diabetes Association.

The author has no conflicts of interest to disclose.

Correspondence: Frank Q. Nuttall, MD, PhD, Minneapolis VA Medical Center, One Veterans Dr 111G, Minneapolis, MN 55417 ( [email protected] ).

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License, where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially.

The body mass index (BMI) is the metric currently in use for defining anthropometric height/weight characteristics in adults and for classifying (categorizing) them into groups. The common interpretation is that it represents an index of an individual’s fatness. It also is widely used as a risk factor for the development of or the prevalence of several health issues. In addition, it is widely used in determining public health policies.The BMI has been useful in population-based studies by virtue of its wide acceptance in defining specific categories of body mass as a health issue. However, it is increasingly clear that BMI is a rather poor indicator of percent of body fat. Importantly, the BMI also does not capture information on the mass of fat in different body sites. The latter is related not only to untoward health issues but to social issues as well. Lastly, current evidence indicates there is a wide range of BMIs over which mortality risk is modest, and this is age related. All of these issues are discussed in this brief review.

Body fatness has been an important psychosocial issue among humans for millennia. It is clearly manifested by paleolithic statuettes of exceedingly plump women. This suggests being “full figured” was highly desirable at least for women. In contrast, images of obese people, males or females, are never exhibited in ancient Egyptian funerary wall paintings, stellae, or statues suggesting that fatness was not considered to be a desirable trait there. This also is the case in artifacts from other cultures in the Middle East in that era. Why the degree of fatness has varied in different cultures is not clear. However, it may have depended on the availability of a reliable food supply and the effort required in obtaining it.

More recently, the degree of rotundity considered ideal also has varied considerably in the general population, but particularly for young women. Before the 1920s, “full figured” women were considered to be desirable as long as the distribution was hourglass in type. However, the 1920s Flapper era introduced abbreviated and revealing dresses. The result was that thinness was not only desirable but also required. This concept has moderated but still influences women’s views of beauty and eating habits at present.

Fatness as a Personal or Society Issue

Traditionally, a person’s fatness has been defined at a personal level as well as at a societal level. However, this is difficult to quantify. That is, each individual has his/her own perception of how fat he/she should be. As indicated above, this often depends on a general concept of societal norms or is due to peer pressure. For example, currently in Western societies, young women are often concerned about their body image, and most consider themselves to be too fat, even though they are well within population-based references. This is not only due to societal concepts of an ideal degree of fatness, but also due to thinness being a goal promulgated by the fashion industry and reinforced by commercial advertising.

At a societal level, although poorly described or quantified, there also is a degree of fatness beyond which a person generally is considered to be unacceptably fat; that is, there is an ill-defined threshold at which a person is labeled as being “fat” or “obese.” However, it is based on the “I can’t define it but I know it when I see it” concept. In addition, implicit in this context is that the location of the excess fat plays a role, as does a person’s age. It is much more acceptable to be “overweight” when one is old than when one is young. Also particularly in women, the accumulation of fat in certain areas of the body is considered to be much more acceptable than in other areas. For example, truncal (belly fat) accumulation would be considered to be less acceptable than the accumulation of fat in the peripelvic and thigh areas as well as in the breast area 1 ; that is, one may be statistically “fat” but with an appropriate figure be merely referred to “as amply endowed” or “pleasingly plump.”

The social consequences of being “too fat” are severe. Discrimination begins in childhood and results in serious emotional scars. Societal discrimination limits career choices, and indeed many career paths are closed to those considered to be too fat. Also, societal stigmatization often impairs a person’s ability to express his/her intellectual and other talents; that is, they become underachievers. In addition, the potential pool of mates is limited because of their perceived unattractiveness. Thus, obese people tend to marry other obese people and, parenthetically, to produce obese children. 2–4

Fatness as a Medical Issue

Not only the societal but also the functional and indirectly the medical consequences of an excessive accumulation of fat also have been recognized for millennia. Nevertheless, the concept that “body build” (fatness) is a major population-based medical issue gained popularity in this country only shortly before 1900. Life insurance data accumulated at that time 5 and subsequently 6 indicated that body weight, adjusted for height (Wt/Ht), was an independent determinant of life expectancy, and in 1910, the effects of being overweight were noted to be greater for younger people than for the elderly. 6

Subsequently, the Metropolitan Life Insurance Company in 1959 published tables of average body weights for heights (Wt/Ht) by gender and at different ages. 7 This was based on data from 1935 to 1953 from more than 4 million adults, mostly men, insured by 26 different insurance companies. The risk for development of certain diseases as well as mortality data related to Wt/Ht differences also were analyzed and reported in the 1960 Statistical Bulletin of the Metropolitan Life Insurance Co. 8,9

The Wt/Ht tables were used for many years as a reference for population-based studies. If a person’s Wt/Ht was 20% above or below the mean for that height category, he/she was considered to be overweight or underweight, respectively. The insurance data also indicated the ratios of weights for heights (the term used was “body build”) at which mortality was lowest in adults. The latter was referred to as the “ideal” or later the “desirable” weight. All of these data were periodically updated. 10 Interestingly, from 1959 to 1983, the desirable weight, that is, the weight/height representing the lowest mortality had increased. 10–12 However, a “desirable body” weight for height was invariably lower than the average weight for height in the insured population. 7,10,13

Problems With the Wt/Ht (Body Build) Index

Early on it was recognized that tall people had a lower death rate than did short people 7,8,13 with the same Wt/Ht ratio. It also was recognized that a person’s height in general and leg length in particular could affect the calculated body mass adjusted for height. A person’s bony frame, that is, bone mass, also could affect the interpretation of this ratio. In general, it reflected whether one was narrowly or broadly built. Thus, efforts were made to eliminate lower limb length and frame size as variables. 7,10 The strategy was to develop representations of body build, that is, charts of weight/height that were independent of these variables. The overall goal was to have the same distribution of Wt/Ht at each level of height.

Although not stated, the implicit goal in developing these tables was to define a person’s fat mass as a proportion of their total mass, irrespective of their height or frame size. 14 Efforts were made to adjust for frame size (nonfat mass) by categorizing people as those with a small, medium, or large frame. Estimation of frame size was attempted using a number of measurements including shoulder width, elbow width, knee width, ankle width, and so on. 15 None of these were widely adopted. Nevertheless, frame size based on elbow width was included in the Metropolitan Life weight/height tables, 7,10 even though it was never validated.

Mathematical Adjustment of Body Build

Mathematically, the issue of adjusting body build for differences in height was approached with the concept that the body, particularly the trunk, could be considered as being a 3-dimensional volume or mass. Thus if a tall person were simply a scaled-up version of a short person, weight would increase approximately with the cube of height. 16 Indeed, several equations were developed and tested based on this concept; that is, the cube root of weight divided by height ( 3 √Wt/Ht) or weight/height, 3 and so on, but none were ideal. 17 This is because tall people are not just scaled-up versions of short people. As indicated previously, they tend to be more narrowly built resulting in a greater lean/fat proportion of body mass.

Later, it was shown that the body mass for height actually scaled best with weight for height when the height was raised to the 1.6 to 1.7 exponent (Wt/Ht, 1.6 etc). 18 Thus, with an increase in Ht, the effect of Ht on the ratio is exponential, whereas the change in Wt is linear. This has the effect of Ht on the ratio to be magnified as Ht increases. Overall, it results in a lower ratio in tall people than will be obtained with just a Wt/Ht ratio. Thus, it potentially compensates for a narrower build in tall compared with short people.

This exponent is not convenient for use in population-based studies, and it was determined that Wt/Ht 2 generally was satisfactory. 16,18 The latter represents the Quetelet Index. It was developed by Dr Quetelet in the 1800s.

Lambert Adolph Jacque Quetelet

I would like to briefly mention who Dr Quetelet was and how the “Quetelet Index” was derived. 19–21 Lambert Adolphe Jacque Quetelet (1796–1874) was a Flemish astronomer and statistician. Indeed, he is considered to be the patriarch of statisticians. He introduced the concept of “social averages.” In developing the “social average” concept, his goal was to determine the characteristics of an “average man” and the distribution of various human characteristics around the “average man.” Overall, it was his desire to obtain a distribution such that it formed a bell-shaped curve, that is, a Gaussian or normal distribution. He referred to his studies as “social physics.” Thus, this represents the first application of distribution mathematics to human characteristics. In 1835, Quetelet noted the body mass relationship to height in normal young adults was least affected by height when the ratio of weight to height squared was used rather than merely using the ratio of the weight to height or weight to height raised to the third power. 16

Adoption of the BMI as an Index of Obesity

In 1972, Keys et al 16 severely criticized the validity of Metropolitan Life Insurance published data per se, and the then-published tables of desirable weight for height, as well as the tables used to define people who were underweight or overweight. 7 (The pejorative term “obese” was rarely used in that era.) Instead, Keys et al, using better documented weight for height data, popularized the Quetelet Index in population-based studies. They referred to it as the body mass index (BMI). Thus, Quetelet Index = body weight (kilograms) divided by height squared (meters) = BMI.

As indicated above, by squaring the height, it reduces the contribution of leg length in the equation and tends to normalize the body mass distribution at each level of height; that is, it reduces the effect of a variance in height in the relationship of weight to height. This was considered to be important because most of body fat is in the trunk. Nevertheless, as also pointed out by Keys et al, 16 even the BMI rather poorly represents a person’s percent of body fat.

Despite all the criticisms, the Metropolitan Life Tables criteria for defining obesity were widely used in the United States until the early 1990s. 22–24 At about that time, the World Health Organization (WHO) classification of body weight for height, based on the BMI, was published, 25 and later it was widely adopted. 26

BMI Distribution in a Normal Population

Although a BMI determination reduces the effect of lower-extremity length on the Wt/Ht ratio, whether one uses the BMI or merely the ratio of weight to height, the population distribution is still not Gaussian. That is, it is not symmetrical but is always skewed to the right, that is, toward a higher ratio of weight (body mass) to height. For example, the distribution of BMIs in adult American men and women was determined in 1923 in 1026 individuals (Figure). 27 The median BMI was 24, but the mean BMI was 25. The distribution curve clearly indicated a skewing toward an increase in BMI, and this trend has continued. 26

This skewing is not surprising because a markedly reduced BMI, theoretically and actually, would be incompatible with life because of an excessive reduction in lean as well as fat mass as a result of under nutrition 28 or disease. In contrast, excessive accumulation of body fat with maintenance or usually an increase in lean mass 29,30 is at least compatible with life, even though it may eventually affect long-term survival.

WHO and the Categorization of BMIs Into Quartiles

In 1993, the WHO assembled an Expert Consultation Group with a charge of developing uniform categories of the BMI. The results were published as a technical report in 1995. 25 Four categories were established: underweight, normal, overweight, and obese. An individual would be considered to be underweight if his/her BMI was in the range of 15 to 19.9, normal weight if the BMI was 20 to 24.9, overweight if the BMI was 25 to 29.9, and obese if it was 30 to 35 or greater. Using linear regression, a BMI of 16.9 in men and 13.7 in women represents a complete absence of body fat stores. 31

The above 4 categories are similar to those suggested by John S. Garrow in 1981, 31,32 but the terminology was changed. The terminology he used was “desirable” for a BMI up to 25, “grade I obesity” between 25 and 29.9, “grade II obesity,” between 30 and 40, and “grade III obesity” for BMI greater than 40.

The latter classification was based on Rosenbaum and colleagues’ 33 own data obtained in a survey of an adult population, aged 16 to 64 years, in Great Britain and published in 1985.

The population-based data indicated the majority of people were in the “desirable” range of the BMI distribution as indicated in Table 1 . Unfortunately, this distribution is not and has not been similar to those found in other surveys. The BMIs have been higher.

At the time that the WHO classification was published, the National Institutes of Health (NIH) in the United States classified people with a BMI of 27.8 (men) and 27.3 (women) or greater as being overweight. If they were below this BMI, they were considered to be “normal.” This was based on an 85% cutoff point of people examined in the National Health and Nutrition Examination Study (NHANES) II. 12,22,34 Subsequently, in 1998, the cutoff point between normal and overweight was reduced to a BMI of 25 to bring it into line with the 4 categories in the WHO guidelines. 25,35 Parenthetically, this instantaneously converted millions of Americans from being “normal weight” to being “overweight.”

In 1997, the International Obesity Task Force expanded the number of BMI categories to include different degrees of obesity and changed the terminology modestly. 36 A BMI of 25 to 29.9 is referred as “preobesity,” a BMI of 30 to 34.9 is class I obesity, 34.9 to 39.9 is class II obesity, and a BMI of 40 or greater is class III obesity. 37,38

The new terminology appears to be a bit presumptuous and careless because the BMI is not a direct measure of percent of fat mass, and the dynamic concept that those in the former “overweight” category are now in the “preobesity” category invariably going on to “obesity” is not the case. Also those with a lower BMI initially, but with a dynamic weight gain over time, would have to transition through this category in order to become classified as “obese” regardless of the terminology. By analogy, should those classified as “underweight” now be referred to as being “prenormal”?

Distribution of BMI in the General Population

It should be understood that in Western population-based studies, generally the mean or median BMI is about 24 to 27. 22,27,39,40 Thus, the consequence of adopting the WHO classification is that ~50% or more of the general adult population will always be in the overweight (now preobese) and obese categories. Indeed, the term “overweight” or particularly “preobesity” is prejudicial since people in this category are a major part of the expected normal distribution of BMI in the general population, and this has been the case for decades. Unfortunately, in discussing the so-called “obesity epidemic,” the number of people in the overweight (preobese) category generally is lumped together with those in the obese category in order to advertise and dramatize the perceived seriousness of this issue.

Regardless of the terminology and population reference issues, at present the BMI is the currency by which we define the obesity issue throughout the Western world. It was developed for the convenience of the epidemiologists, and indeed it did provide a uniform codification of body weight for height reporting. The BMI categories are shown in ( Table 2 ).

BMI as a Determinant of Body Fat Mass

A particular problem with BMI as an index of obesity is that it does not differentiate between body lean mass and body fat mass; that is, a person can have a high BMI but still have a very low fat mass and vice versa. 39,41–46

From an anatomical and metabolic perspective, the term obesity should refer to an excessive accumulation of body fat (triacylglycerols), and upon these grounds, the accuracy of the BMI as a determinant of body fat mass has been repeatedly questioned, 16,39–41,46–48 because it clearly has limitations in this regard. Gender, age, ethnic group, and leg length are important variables. 45,49–55 It should be noted that in population-based studies women generally have a BMI that is lower than that in men, even though their fat mass relative to their body build or BMI is considerably greater (~20% to 45%+).

The relatively poor correlation between percent of body fat mass and BMI in males has been known for many years 16 and was clearly shown in a study in which percent of body fat was determined by a densitometric method. 56 For men with a BMI of 27 in that study, the 95% confidence intervals for percent of body fat were 10% to 32%; that is, in this group, the percent of body fat varied from very little to that considered to be in the obesity range. (NIH-suggested criterion for obesity based on percent of body fat for men is ≥25%, and that for women is ≥35%. 57 )

The relatively poor correlation between percent of body fat mass and BMI also clearly has been shown more recently in the NHANES III database in which bioelectrical impedance was used to estimate the fat component of body composition. 51 In subjects with a BMI of 25 kg/m 2 , the percent of body fat in men varied between 14% and 35%, and in women it varied between 26% and 43%. Thus, using the NIH-suggested criterion based on percent of body fat to define obesity, subjects with a BMI of 25, a group that would be considered to be essentially normal, were associated with a body fat mass that varied again between low normal to obese. Also it is of interest that in the entire NHANES cohort, the BMI correlated better with lean body mass than with fat mass in men. 51 More recent NHANES data also indicate a poor correlation of BMI with percent of body fat, particularly in men. 58

In addition, in a recent study in individuals with or without diabetes in which the loss of lean body mass with aging was reported, the mean BMI in those without diabetes was 26.8. In those with diabetes, the BMI was 29.1; that is, it was higher as generally expected. However, the percent of lean body mass was the same; that is, the increased BMI in those with diabetes was not due only to an excessive accumulation of fat. 59

Trends in Body Weight and Height

Over the past several decades, there has been an increase in BMI in the general population. This has resulted in predictions of a public health disaster. It should be recognized that in the United States during the period from 1960 to 2002 not only has the mean weight increased by 24 lb among men aged 20 to 74 years, but also the height has increased by about 1 in. We can then calculate that the weight increase per year has only been 0.57 lb, and as indicated above, this could be due to an increase in lean mass rather than fat mass, or it may be a combination of the two. In women, there was a similar increase in weight and height. 40

In an earlier report, life-insured men up to age 40 years were reported to be 0.5 to 1.5 inches taller and 2 to 9 lb heavier for the same height in 1959 than those studied 50 to 60 years prior to 1959. Also, in the earlier study, the mortality rate was lowest in those with higher weight-to-height ratios. This was attributed to the presence in the population of wasting diseases such as tuberculosis that resulted in an increased death rate. 13 Previously, a secular upward trend in height in adults in the United Kingdom also was reported. 60 In addition, in a twin study in the United Kingdom, children in 2005 were not only heavier but also taller than 1990 norms, whereas their BMIs were essentially the same. 3

Overall, the history of changes in height and weight in Western European men and probably women has been that of an increase in both weight and height. In the 17th century, the average height of men in Northern Europe was ~5 ft 3 in. It now approaches 6 ft. 61 These data suggest that the BMI categories should be adjusted upward periodically to accommodate population-based changes. Improvements in mortality rates also suggest an adjustment would be useful.

Body Fat Location

An additional limitation of the BMI is it does not capture body fat location information. This is an important variable in assessing the metabolic as well as mortality consequences of excessive fat accumulation. It was first recognized in France by Dr Jon Vague 62 in the 1940-1950s. He noted that accumulation of fat in the upper part of the body versus the lower part of the body was associated with an increased risk for coronary heart disease, diabetes, and also gallstones and gout. That is, individuals who accumulated excessive fat in the lower body segment were relatively spared from these complications. The body fat distribution was referred to as being “android” if it occurred in the upper body and “gynecoid” when it occurred in the lower segment of the body. This is because men tend to accumulate fat in the abdominal (upper body) area, whereas women tend to accumulate it in the peripelvic (gluteal) area and the thighs. A surrogate for this information has been to determine the abdominal circumference or an abdominal/hip circumference ratio. Subsequent data indicate that indeed the risk for development of diabetes and the so-called “metabolic syndrome,” as well as coronary heart disease, is more strongly related to the accumulation of upper body fat than lower body fat in both sexes. 63–67 That is, an android (male) distribution more closely predicts the development of the chronic diseases of aging than does the gynecoid (female) distribution.

More specifically, both visceral fat accumulation 68,69 and an expanded girth have been associated with development of insulin resistance, diabetes, and risk for coronary heart disease and hypertension. 63,64,70–74 Accumulation of fat in the abdominal area appears to correlate best with triacylglycerols accumulating in the liver and skeletal muscle. These may actually represent the pathogeneticially important metabolic consequences that result in insulin resistance and more directly correlate with development of the above adverse medical conditions. 68,75,76 Incidentally, the relatively small accumulation of fat in these organs would not be detectible by BMI determinations, and they do not correlate simply with total body fat mass. 75

The Life Cycle and Location of Accumulated Fat

Prior to puberty, boys and girls tend to be lean and not much different in this regard. Girls tend to accumulate relatively large amounts of fat during and after puberty, particularly in the peripelvic and thigh region; boys do not. During and after puberty, boys accumulate a relatively large amount of lean mass (bone and muscle) but not fat mass. In both sexes, these changes are reflected in an increased BMI. With aging, both sexes tend to develop fat in the upper part of the body (circumferentially), that is, the middle-age spread. 49,77–80 The reason for these changes in amount and distribution is not completely understood. Generally, it is considered to be due to hormonal changes.

It is of some interest that accumulation of fat in the lower body at puberty in females is unique to humans, is not present in any of the great apes, and most likely is estrogen mediated. 1

In a teleological sense, why this occurs, if due to estrogen, is uncertain. It could represent a means of maintaining body fat during pregnancy without an undue expansion in abdominal girth. It also may act as a counterbalance when women carry a child either during pregnancy or afterward. It also may be a space-filling fat site due to the relatively larger pelvis in postpubertal females. 81 Overall, it may represent an adaptation to the human upright bipedal posture. In any event, it results in a lower center of gravity among women compared with men. Indeed, the lower body segment in females becomes ~40% greater than in males (quoted in Singh, 1993), 1 and it has strong sex-related overtones.

Not only is thigh fat greater in women than in men, but also women generally have a preponderance of slow-twitch fibers, whereas men have a preponderance of fast-twitch fibers in their quadriceps muscles, as do upper-body-obese women, 82 suggesting either genetic or earlier developmental differentiation events. Could this be an adaptation for load-bearing versus speed as a group survival adaptation?

As indicated above, the accumulation of fat with aging in both sexes tends to occur in the truncal area and is associated with an increase in visceral fat. In women, this could be explained by a decrease in circulating estradiol, that is, a decreased estrogen/testosterone ratio associated with the menopause. (Again of some interest, it is only humans who have a defined menopause).

In men, with aging, there is a decrease in testosterone and a relative increase in estrogen, resulting in a decrease in the testosterone/estrogen ratio. 83 Thus, in men, a change in sex hormone concentrations could possibly explain the increased accumulation of fat in general. However, why there is a preferential accumulation in the truncal location, that is, why they too develop an increase in visceral fat, is unclear. Clearly, location of fat in this area would help to maintain mobility. The latter could be of great importance in hunter-gatherer societies and in defense of the tribe. Perhaps the distribution is programmed by gender earlier in life.

In this regard, it should be recognized that the accumulation of fat in certain body areas as well as the total amount of fat accumulated has a strong genetic or at least a familial component that diminishes with age. 3,27,84,85

Methods of Estimating Body Fat Mass and Location of the Fat

At present, simple, accurate methods for measuring percent of body fat and, in particular, body fat in different fat depots are not available. The indirect methods currently in use for estimating total percent of body fat include underwater weighing, an air displacement and density determination using a Bod Pod, a bioelectrical impedance analyzer, and a determination of the isotopically labeled water mass. In the past, determination of the total body radioactive potassium and thus metabolizing tissue mass have been used to estimate lean body mass, and by difference, the fat mass. 86

Anthropometric determination of fat mass directly has been done using skin-fold thickness measured at various sites. 87 A dual-energy x-ray absorptiometry (DEXA) scan, which provides a 3-dimensional picture of body organ densities, can be used for estimating total body fat. Its location also can be determined. Single computed tomography (CT) slices of the abdomen and thigh can be used to obtain 2 dimensions of those fat depots from which a 3-dimensional fat area can be reconstructed. This also can be done using magnetic resonance imaging, but magnetic resonance imaging is very expensive. One cannot do serial sections of the body using CT to determine fat mass because of the excess radiation associated with this procedure.

Because of their convenience, bioelectric impedance methods or DEXA scans are the most commonly used to estimate the amount and, with DEXA scans, the location of body fat depots. Estimates of abdominal and thigh fat depots also can be estimated using CT slices. 52,72,88

All of the previously mentioned methods use certain assumptions in the calculation of body fat mass, and all are subject to potential error. Nevertheless, there are more specific methods of determining body fat mass than is the BMI. Important information regarding the location of the stored fat also can be determined with some methods.

It now is generally accepted that a relationship between BMI and mortality risk should be applied only to large populations. It should not be applied to an individual in an unqualified fashion. As indicated previously, there is the issue of being “overweight” versus “over fat.” In addition, a segment of the population is now considered to be “fat” by any criteria but “fit” and not at risk for early mortality. 74,75,89–91

BMI and Morbidity and Mortality

The BMI classification system currently is being widely used in population-based studies to assess the risk for mortality in the different categories of BMI. It also is being used in regard to specific etiologies for mortality risk. However, as with its use to estimate percent of body fat, it is a rather crude approach. Even when some comorbidities are considered, the correlation of mortality rates with BMI often does not take into consideration such factors as family history of diabetes, hypertension, coronary heart disease, metabolic syndrome, dyslipidemias, familial longevity or the family prevalence of carcinomas, and so on. Recently it has been reported that more than 50% of susceptibility to coronary artery disease is accounted for by genetic variants. 92

Frequently, when correlations are made they also do not take into consideration a past as well as a current history of smoking, alcohol abuse, serious mental disorders or the duration of obesity, when in the life cycle it appeared, and whether the body weight is relatively stable or rapidly progressive, that is, type 1 or type 2 obesity. 93,94 In most population-based studies, only the initial weight and/or BMI are given, even though weight as well as fat stores are known to increase and height to decrease with aging. In addition, the rate of weight gain varies among individuals, 7,94,95 as does the loss of muscle mass. 95 Muscle mass has been correlated negatively with insulin resistance and prediabetes. 96 Lastly, population-based studies do not take into consideration the present and past history of a person’s occupation, medication-induced obesity, and how comorbidities are being treated. All of the above are significant issues.

More Explicit Problems in Relating the BMI to Medical Issues

Based on data in the literature, there are several additional problems in determining associations between BMI and overall death rate or, more specifically, cardiovascular events or death rates. Many obese people do not have cardiovascular risk factors, and in those who do, BMI no longer correlates with cardiovascular events 97–101 when the untoward effects of these other factors are factored out. Another issue is that the treatment status of the previously mentioned cardiovascular risk factors often is unknown or not stated; that is, the efficacy of treatment is rarely considered. This also is the case for diabetes. For example, the prevalence of diabetes has been increasing but not the disease-specific death rate. 102 Also, in people with diabetes, the death rate from cardiovascular disease has decreased dramatically. 102

The “Obesity Epidemic”

Recently, there has been concern that an epidemic of obesity is occurring, not only in the United States, but also worldwide based on BMI data. In the NHANES data, there has been a change in the mean but to a greater extent in the distribution of BMI for adults aged 20 to 74 years in the United States. 26 That is, the mean BMI has increased, but there has been a greater increase in skewing toward the right and very large BMI. This results in more individuals being categorized as “obese.” The reason for the recent increase in mean BMI, but particularly in those in the obese category, is unknown, although there are many speculations. The dramatic decrease in smoking is likely to have been a contributor. 91,103–106 Smoking contributes to population-based BMI by at least 2 mechanisms. Smoking impairs appetite per se. It also is pathogenetically important in the development of chronic obstructive pulmonary disease, which itself results in a lower body mass. Of some interest, NHANES data also indicate that the trend of an increase in BMI has not continued since 1999 in women and only modestly in men. 58 Smoking rates also have stabilized at a low level.

Is Being “Overweight” by BMI Criteria a Medical Issue?

Regardless of an observed increased skewing in the BMI distribution, it is important to note that several recent studies indicate that for most of us being a bit overweight (preobese?) as determined by BMI may not be so bad. 107–111

The EPIC observational study is a population-based study that includes 359 387 individuals aged 25 to 70 years living in Europe. 109 The mean age of this group at the initiation of the study was 51.5 years, and the mean follow-up has been 9.7 ± 2 years. In this study, both the crude and adjusted relative risk of death among men was actually the lowest in those with a BMI of 26.5 to 28, that is, those in the overweight (preobese) category. Also, a significant increase in risk of death was present only among those with a BMI of less than 21 or greater than 30. That is, there is a wide range of BMIs in the central part of this population in which there was relatively little impact of BMI on risk of death over a 9.7-year period.

Similar data were obtained in the NIH–American Association of Retired Persons study of 527 265 men and women between the ages of 50 and 71 years in the United States and followed for up to 10 years. 110 The lowest death rate in the entire cohort was among those in the “overweight” category, and this was particularly true among the men. There also was a broad range of BMIs over which there was little difference in mortality (BMI of 23.5 to 30).

The NHANES data going back to 1971 and up to 1994 also indicate that the relative mortality risk is lowest in men with a BMI of 25 to 30 in all age groups, that is, from the age of 25 years up to the age of 70 years. 107 In addition, the risk of mortality was little affected by a BMI from 18.5 up to a BMI of 30 in all age categories. Indeed, in those older than 70 years, there was little impact on the death rate even if they were in the obese category. Similar results have been reported for women in the NHANES reports. 112 The lowest mortality occurred with a BMI of 27.

In a Canadian study, the age-adjusted mortality rate over 13 years in men was essentially unchanged in those with a BMI of 18.5 up to 35, that is, from the Normal Weight category through the obesity class I category. In women, there was only a modest increase over the same range. 113

In summary, there is a large range of BMIs over which there is little association with the death rate. Generally, the range is from a BMI of 21 up to and often including 30. It is centered in the 24-to-28 BMI range. This information is not entirely new. Andres 114 in 1980 summarized 16 different population-based studies in which anthropometrically determined obesity was not associated with increased mortality rate. A detailed analysis in 1960 of the Metropolitan Life Insurance data also suggested little increase in mortality rates in people with a degree of overweight less than 20% or more above the average for a given height and age (quoted in Keys et al 97 ).

Interestingly, in the EPIC observational Study, 109 when the waist circumference–to–BMI ratio was calculated, that is, adjusting the waist circumference for BMI, it tended to linearize the association of BMI with risk for death, and the ratio was greatest for those with a low BMI. Thus, even if an individual had a low BMI but a relatively increased waist circumference, the risk was increased. Indeed, for any given BMI, a 5-cm increase in circumference increased the risk of death by a factor of 1.17 among men and 1.13 for women. Also in this study, the overall greatest mortality risk was in those individuals with the lowest BMI and not those with the highest BMI. Nevertheless, even in the category with the lowest BMI, adjusting for waist circumference affected the mortality rate negatively. This again indicates the importance of the location of body fat in addition to the total amount of fat accumulated.

A recent analysis of 50 prospective observational studies indicated the lowest mortality at a BMI of 23 to 25. However, these data were obtained in the 1970s and 1980s in an aggregate population with a mean BMI of 24.8, that is, lower than at present. The increased mortality at higher BMI’s was modest up to a BMI of 27.5, and the authors could account for the excess mortality largely on the risk factors known to be associated with obesity. The latter are currently being much better treated than in that era. 115

Issues to be Resolved When Relating BMI With Health Determinants

Overall, a major unresolved issue is which factor of the following is more important in the prediction of comorbidities such as cardiovascular disease, diabetes, hypertension, malignancies, or overall death rates. Is it BMI, total body fat mass, or the distribution of body fat, that is, visceral versus subcutaneous, or upper body fat accumulation (as determined by abdominal circumference, or a waist/hip ratio, or some combination of these, and so on)? The EPIC 109 data suggest that where fat is accumulated is much more important than merely the BMI, with the exception of those with an exceeding large total fat mass.

SUMMARY AND CONCLUSION

It is time to move beyond the BMI as a surrogate for determining body fat mass. Alternatively, if BMI continues to be used, the categories and definitions should be changed to reflect the current distribution of BMIs in the general population.

A better means than the BMI for estimating percent of body fat and its relationship to mortality and various morbidities clearly would be desirable.

The BMI was not originally developed for use specifically as an index of fatness in population-based studies. However, it has been coopted for this use because it is a readily obtained metric. It should be understood that the BMI has serious limitations when used as an indicator of percent of body fat mass. Indeed, it may be misleading in this regard, particularly in men. The terminology currently used also is prejudicial. By definition, one-half or more adults in the recent past and currently are overweight (preobese) or obese in Western, industrialized nations.

The current BMI classification system also is misleading in regard to effects of body fat mass on mortality rates. The role of fat distribution in the prediction of medically significant morbidities as well as for mortality risk is not captured by use of the BMI. Also, numerous comorbidities, lifestyle issues, gender, ethnicities, medically significant familial-determined mortality effectors, duration of time one spends in certain BMI categories, and the expected accumulation of fat with aging are likely to significantly affect interpretation of BMI data, particularly in regard to morbidity and mortality rates. Such confounders as well as the known clustering of obesity in families, the strong role of genetic factors in the development of obesity, the location in which excessive fat accumulates, its role in the development of type 2 diabetes and hypertension, and so on, need to be considered before promulgation of public health policies that are designed to apply to the general population and are based on BMI data alone.

Clearly, obesity, as determined by BMI, is not a monotypic, age-invariant condition requiring a general public health “preventative” approach. A BMI-determined categorization of an individual should not be used exclusively in counseling or in the design of a treatment regimen. In addition, when considering weight loss regimens, variations in body weight attributed to weight loss and dietary cycling may be hazardous. 116–120 They have been associated with an increased mortality rate. 116,117,119,121–124 The concept of starvation-associated obesity 125,126 also needs to be considered.

Prevention and/or Treatment of BMI-Determined Overweight or Obesity

Clearly episodic starvation or semistarvation regimens are not the answer, 127 nor are population-based efforts to increase fresh fruits and vegetables and tax soda pop, and so on. In my opinion, the major focus on prevention and treatment should be on those unfortunate individuals who are grossly obese, mechanically compromised, and who are at very high risk for death. 128 Surgical gastrointestinal intervention has proven to be at least partially successful in improving fuel regulation and storage. 129,130 Hopefully, medications will be developed that will reinstitute a metabolic fuel regulatory system that prevents the relentless accumulation of body fat, which is characteristic of those who are grossly obese. For others, an improvement in physical fitness may be salutary.

A Personal Perspective Regarding the Obesity Epidemic

Currently there are 4 truths regarding historical changes in body weights and the prevalence of obesity. People of Western European extraction are on average (1) heavier, (2) taller, and (3) more likely to be “overweight” or “obese” as defined by current BMI standards than those in other parts of the world. However, (4) it also should be pointed out they are healthier and are living longer than in any previous period in history. 131,132

Beginning in the 17th century, 61 the general underlying theme in all the studies done on weight gain in populations is an increase in height as well as weight. These changes are likely to be due to an increase in high-quality dietary protein (animal products), as well as an increased availability of total food energy in the diet. That is, there was not only an increase in food availability and variety, but also an increase in food quality. 133 The near elimination of chronic and serious acute infectious diseases also may have played a role, as has the dramatic decrease in cigarette smoking and its serious medical consequences.

The net effect of the above is that the chronic diseases of aging have become more of a public health problem, but better treatments are widely available. The prevalence of type 2 diabetes has increased, but overall the cardiovascular death rate has decreased dramatically. The death rate from malignancies is decreasing, and there has been a remarkable improvement in longevity, which is continuing. 131 The latter also is likely to continue into the future. 131,132

Some view the secular trend in the US population over the past 40 years as being one in which the population in general is “more obese, more diabetic, more arthritic, more disabled, and more medicated” but living longer. 134 A less sanguine view is indicated by others. 135 Many consider the overabundance of “calorie dense, processed foods,” the availability of soda pop, 136 and presence of fast-food restaurants and large food portion sizes to be strong, pathogenetic, obesity-inducing factors, 137 or more broadly, they consider obesity to be due to a “toxic” or “poisonous” food supply. 138 Some also are concerned that the increase in obesity (defined by BMI) will overwhelm any gains in health and life expectancy noted over the past several decades, that is, an Apocalypse awaits us. 139 I and others 140,141 do not share this pessimism.

Finally, I would like the political activists and doomsday prophets whose professional careers appear to depend on frightening the public and inducing politicians to pass restrictive laws without proven value, to be introduced to the prescient comments made by A. E. Harper 133 33 years ago. It is clear that currently we have a case of “déjà vu all over again.”

In regard to predicting the future, a wise person whose name I cannot recall stated presciently “Predicting the future is a fool’s playground”; the physicist Neils Bohr said, “Prediction is very difficult, especially about the future,” or as stated by that sage of the baseball world, Yogi Berra, “The future ain’t what it used to be.” Bertrand Russell said, “Fools and fanatics are always so sure of themselves, but wiser people are so full of doubt.” The true scientist should always be a skeptic.

Acknowledgments

The author thanks Rachel Anderson for help in preparing the manuscript for submission and Dr Mary C. Gannon for reading the manuscript and making numerous helpful comments.

• Cited Here |
• PubMed | CrossRef
• View Full Text | PubMed | CrossRef
• + Favorites
• View in Gallery

Readers Of this Article Also Read

Remaining relevant in a changing healthcare organization: registered dietitian..., nutrition 2024 registration now open, pistachios for health: what do we know about this multifaceted nut, rosemary: an overview of potential health benefits, raw milk consumption: risks and benefits.

Margarets bmi is 283 read principles of nutri on sec

• Brigham Young University, Idaho
• NUTRITION 150
• KidViper1025

End of preview

Recently submitted questions.

• Research article
• Open access
• Published: 29 January 2024

Integrating multiple lines of evidence to assess the effects of maternal BMI on pregnancy and perinatal outcomes

• Maria Carolina Borges 1 , 2   na1 ,
• Gemma L. Clayton 1 , 2   na1 ,
• Rachel M. Freathy 1 , 3 ,
• Janine F. Felix 4 , 5 ,
• Alba Fernández-Sanlés 1 , 2 ,
• Ana Gonçalves Soares 1 , 2 ,
• Fanny Kilpi 1 , 2 ,
• Qian Yang 1 , 2 ,
• Rosemary R. C. McEachan 6 ,
• Rebecca C. Richmond 1 , 2 ,
• Xueping Liu 7 ,
• Line Skotte 7 ,
• Amaia Irizar 8 , 9 , 10 ,
• Andrew T. Hattersley 3 ,
• Barbara Bodinier 11 ,
• Denise M. Scholtens 12 ,
• Ellen A. Nohr 13 ,
• Tom A. Bond 1 , 2 , 11 , 14 , 15 ,
• M. Geoffrey Hayes 16 ,
• Jane West 6 ,
• Jessica Tyrrell 3 ,
• John Wright 6 ,
• Luigi Bouchard 17 ,
• Mario Murcia 10 , 18 ,
• Mariona Bustamante 10 , 19 , 20 ,
• Marc Chadeau-Hyam 21 ,
• Marjo-Riitta Jarvelin 21 ,
• Martine Vrijheid 10 , 19 , 20 ,
• Patrice Perron 22 , 23 ,
• Per Magnus 24 ,
• Romy Gaillard 4 , 5 ,
• Vincent W. V. Jaddoe 4 , 5 ,
• William L. Lowe Jr 16 ,
• Bjarke Feenstra 7 ,
• Marie-France Hivert 25 , 26 ,
• Thorkild I. A. Sørensen 27 , 28 ,
• Siri E. Håberg 24 ,
• Sylvain Serbert 29 ,
• Maria Magnus 24 &
• Deborah A. Lawlor 1 , 2  

BMC Medicine volume  22 , Article number:  32 ( 2024 ) Cite this article

2967 Accesses

2 Citations

94 Altmetric

Metrics details

Higher maternal pre-pregnancy body mass index (BMI) is associated with adverse pregnancy and perinatal outcomes. However, whether these associations are causal remains unclear.

We explored the relation of maternal pre-/early-pregnancy BMI with 20 pregnancy and perinatal outcomes by integrating evidence from three different approaches (i.e. multivariable regression, Mendelian randomisation, and paternal negative control analyses), including data from over 400,000 women.

All three analytical approaches supported associations of higher maternal BMI with lower odds of maternal anaemia, delivering a small-for-gestational-age baby and initiating breastfeeding, but higher odds of hypertensive disorders of pregnancy, gestational hypertension, preeclampsia, gestational diabetes, pre-labour membrane rupture, induction of labour, caesarean section, large-for-gestational age, high birthweight, low Apgar score at 1 min, and neonatal intensive care unit admission. For example, higher maternal BMI was associated with higher risk of gestational hypertension in multivariable regression (OR = 1.67; 95% CI = 1.63, 1.70 per standard unit in BMI) and Mendelian randomisation (OR = 1.59; 95% CI = 1.38, 1.83), which was not seen for paternal BMI (OR = 1.01; 95% CI = 0.98, 1.04). Findings did not support a relation between maternal BMI and perinatal depression. For other outcomes, evidence was inconclusive due to inconsistencies across the applied approaches or substantial imprecision in effect estimates from Mendelian randomisation.

Conclusions

Our findings support a causal role for maternal pre-/early-pregnancy BMI on 14 out of 20 adverse pregnancy and perinatal outcomes. Pre-conception interventions to support women maintaining a healthy BMI may reduce the burden of obstetric and neonatal complications.

Medical Research Council, British Heart Foundation, European Research Council, National Institutes of Health, National Institute for Health Research, Research Council of Norway, Wellcome Trust.

Peer Review reports

Obesity is a leading preventable cause of ill health, mortality, and morbidity across the world and affects 10% and 25% of adult women in low- and high-income countries, respectively [ 1 ]. Higher maternal pre-pregnancy body mass index (BMI) is associated with a higher risk of various adverse pregnancy and perinatal outcomes, including pregnancy loss, gestational hypertension (GH), preeclampsia (PE), gestational diabetes mellitus (GDM), perinatal depression, caesarean deliveries, preterm birth (PTB), large for gestational age (LGA), and no breastfeeding initiation [ 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 ]. However, given the ethical and logistical challenges of conducting randomised controlled trials (RCTs) in pregnancy, most evidence in the field comes from conventional observational studies, which may be confounded by unmeasured or inaccurately measured maternal characteristics, such as socioeconomic position, age, parity, ethnicity, smoking, and alcohol intake.

Understanding the impact of maternal pre-pregnancy BMI on pregnancy and perinatal health is key to inform appropriate interventions aimed at preventing adverse outcomes and to predict their future burden in different populations. A better understanding of the potential causal role of BMI can be achieved by integrating multiple lines of evidence in a triangulation framework [ 13 , 14 ], which can help overcome fundamental biases arising from the reliance on a single method (e.g. multivariable regression in observational studies). In this context, more credible causal inference can be made for findings in agreement across different analytical approaches with different strengths and limitations; while disagreement could decrease confidence in previous findings or highlight specifics of future research needs, for example where there is imprecision in results from some approaches.

The aim of this study was to explore the relation of maternal pre-/early pregnancy BMI (hereafter ‘maternal BMI’) with a wide range of pregnancy and perinatal outcomes by integrating evidence from multivariable regression, Mendelian randomisation, and paternal negative control. The combination of these three approaches provides a unique contribution to the evidence basis on the causal effect of maternal BMI given their different strengths and limitations. While findings from conventional observational studies using multivariable regression might be biased by residual confounding, Mendelian randomisation studies are less prone to such form of confounding but may be biased by weak instruments or unbalanced horizontal pleiotropy [ 15 , 16 ]. The use of negative control designs, such as using paternal BMI as a negative control exposure, can reveal bias in associations of maternal BMI with adverse pregnancy and perinatal outcomes since paternal BMI is unlikely to affect these outcomes, but may be associated with unmeasured confounders in a similar way to maternal BMI (Fig.  1 ) [ 17 , 18 ].

Overview of the three analytical approaches used to investigate the effect of maternal body mass index on adverse pregnancy and perinatal outcomes. A brief description of each approach is presented in the context of exploring the effect of maternal BMI on APPOs’ risk. Given each approach has different strengths and limitations, findings that agree across approaches are likely to be more credible. The description of each approach is simplified for illustration purposes. An extensive description of assumptions and sources of bias for each approach has been reported previously (e.g. [ 17 , 18 , 19 , 20 , 21 ]). The box around the confounders in the multivariable regression reflects the assumption of the method that all confounders were accurately adjusted for in the analyses. BMI, body mass index; APPOs, adverse pregnancy and perinatal outcomes

Study participants

Data were obtained from up to 446,526 women participating in 14 studies in Europe and North America as part of the MR-PREG collaboration [ 22 ] (Table 1 ). We included women who had available information on at least one outcome of interest, had a singleton birth, delivered a baby without a severe known congenital anomaly, and were of European ancestry since most studies included participants of European descent only or predominantly. Informed consent was obtained from all participants and study protocols were approved by the local, regional, or institutional ethics committees. Details of recruitment, data collection, and ethical approval of each study can be found in Additional file 1 : Supplementary Methods [ 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 ].

Exposure measures

Maternal BMI in kg/m 2 was calculated from measured or self-reported weight and height data (Table 1 ). Weight data was collected before pregnancy in eight studies, before 20 weeks of gestation in three studies, and between 24 and 32 weeks of gestation in one study. Two studies did not have a measure of pre- or early-pregnancy BMI and could only contribute to the Mendelian randomisation analyses.

Outcomes measures

We focused on 20 a priori selected (based on clinical relevance and consensus amongst the study team) binary outcomes: miscarriage, stillbirth, hypertensive disorders of pregnancies (HDP), GH, PE, GDM, maternal anaemia, perinatal depression, pre-labour membrane rupture, induction of labour, caesarean section, PTB, LGA, small-for-gestational age (SGA), low birthweight, high birthweight, low Apgar score after 1 min, low Apgar score after 5 min, neonatal intensive care unit (NICU) admission, and breastfeeding initiation (see Table 2 for definitions and total sample sizes). We included related traits amongst the selected outcomes to maximise the number of cohorts contributing to the analyses (e.g. studies that did not have data on gestational age could contribute with information on low birthweight but not SGA). In additional analyses, we examined four continuous traits that underlie some of these outcomes (i.e. birthweight, birth length, ponderal index at birth, and gestational age at birth). Details on outcomes definitions, distributions, and sample sizes for each contributing study are available in Additional file 1 : Supplementary Methods [ 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 ] and Additional file 2 : Supplementary Tables 1A and B.

Covariables

The following were a priori considered potential confounders of the association between maternal BMI and the pregnancy and perinatal outcomes: maternal age, parity, education, smoking during pregnancy, and alcohol use during pregnancy. We also adjusted for offspring sex to improve statistical efficiency given its strong association with some outcomes (e.g. birthweight-related outcomes). Details of the distribution of these covariables in each study are provided in Additional file 2 : Supplementary Table 2.

Statistical analyses

All analyses were conducted using Stata version 17 (StataCorp, College Station, TX) [ 59 ] or R version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria) [ 60 ]. Results are presented as odds ratio (OR) for each binary outcome per standard deviation (SD) increase in maternal BMI to facilitate the comparison of results. The analytical code is available at: https://github.com/gc13313/matbmi_preg .

Multivariable regression analyses

In the main analyses, we used logistic regression with two sets of adjustments: (1) maternal age and offspring sex and (2) additionally maternal education, parity, smoking during pregnancy, and alcohol use during pregnancy where available. We present the fully adjusted model as the main analyses and include the minimally adjusted model in the supplementary material. Similar multivariable linear regression models were used for the additional analyses with continuously measured outcomes. Study-specific results were combined using fixed-effects metanalyses (inverse-variance weighted) for the main analyses assuming that there is one true effect size underlying all included studies, and random-effects metanalyses ( DerSimonian and Laird method ) for sensitivity analyses.

Mendelian randomisation analysis

We used two-sample Mendelian randomisation, in which the effect of interest is estimated by combining summary data for the association of single nucleotide polymorphisms (SNPs) with BMI and with each outcome, as summarised in Fig.  2 [ 61 ]. This approach allowed us to maximise statistical power by including all 14 studies in the analyses even when data on maternal BMI was not available (i.e. FinnGen and UK Biobank).

Overview of the two-sample Mendelian randomisation analyses framework. We selected 97 SNPs as instruments for maternal BMI from a genome-wide association studies (GWAS) metanalysis conducted by the Genetic Investigation of ANthropometric Traits (GIANT) consortium [ 62 , 63 ], including 339,226 males and females. For the selected SNPs, we extracted summary data for the SNP-BMI associations from the GIANT GWAS metanalyses of European ancestry females ( N  = 171,977) and SNP-outcomes associations from European ancestry females from the MR-PREG collaboration ( N range = 92,002 to 446,526). After harmonising SNP-BMI and SNP-outcomes’ summary data, two-sample MR analyses were carried out using the inverse variance weighted (IVW) method, and a series of sensitivity analyses was performed to assess the plausibility of the core Mendelian randomisation assumptions as specified in the figure. For two studies (Generation R and INMA), summary data was only available to us for 32 SNPs reported in an earlier GIANT BMI GWAS [ 63 ], of which 12 SNPs overlapped with the 97 selected SNPs and were included in our metanalyses

We selected 97 SNPs previously reported to be strongly associated with BMI ( P  <  \({5 x 10}^{-8}\) ) from a genome-wide association studies (GWAS) metanalysis conducted by the Genetic Investigation of ANthropometric Traits (GIANT) consortium (Additional file 2 : Supplementary Table 3) [ 62 ]. Unlike more recent BMI GWAS [ 64 ], the cohorts included in this GWAS were largely independent from the studies included in our analyses avoiding potential biases due to sample overlap [ 65 , 66 ].

Summary data for the SNP-BMI associations were obtained from the GIANT GWAS metanalyses of European females (Additional file 2 : Supplementary table 3) [ 62 ], which included up to 171,977 women (~ 0.5% of participants were also included in our study). We estimated the strength of the genetic instruments using the mean F-statistic and total R 2 for the SNP-BMI association in the GIANT GWAS results as previously described [ 67 , 68 ]. We also examined the correlation between SNP-BMI estimates in non-pregnant (data from the GIANT consortium) and pregnant women (data from participating cohorts where information on maternal BMI was available to us).

Summary data for the SNP-outcomes associations were obtained from each contributing study using logistic (or linear) regression assuming an additive model. For each SNP, we meta-analysed cohort-specific SNP-outcome associations using inverse-variance weighted fixed-effects for the main analyses and random effects ( DerSimonian and Laird method ) for sensitivity analyses.

The main two-sample MR analyses were carried out using the inverse variance weighted (IVW) method [ 67 ]. In addition, we also conducted a leave-one-out analysis at the study level where the pooled IVW estimates were re-computed removing one study at a time to check whether pooled results were driven by a single study.

We conducted a series of sensitivity analyses to explore the plausibility of the core Mendelian randomisation assumption that any effect of SNPs on the outcomes is fully mediated by maternal BMI. We explored the potential presence of invalid instruments (e.g. due to SNPs affecting the outcomes through pathways not mediated by BMI) by (i) assessing between-SNP heterogeneity and directional pleiotropy in effect estimates using Cochran’s Q-statistic and the MR-Egger intercept test [ 68 ], respectively, and (ii) using other Mendelian randomisation methods that are more robust to invalid instruments than IVW (MR-Egger [ 68 ], weighted median [ 69 ], and weighted mode [ 70 ]). For offspring outcomes (Table 2 ), we explored whether IVW estimates might be biased by genetic confounding since maternal BMI genetic variants might influence offspring outcomes (e.g. birthweight) due to the foetus inheriting these variants from the mother rather than due to a causal effect of maternal BMI on the intra-uterine environment [ 71 , 72 , 73 ]. This was done by repeating the IVW analyses using summary data for the SNP-outcomes associations adjusted for offspring genotype, which were obtained by regressing each outcome on the maternal genotype for each SNP including the offspring genotype for the respective SNP as a covariable in the model (all genotypes were coded as the number of BMI-increasing alleles).

Paternal negative control analyses

We used paternal BMI as a negative control exposure to explore whether the associations of maternal BMI with pregnancy and perinatal outcomes could be explained by residual confounding due to shared familial environment influencing BMI in both partners [ 18 , 74 ]. These analyses included paternal BMI data from ALSPAC ( N  = 2821–6952), calculated from weight and height self-reported by the father during the first trimester; GenR ( N  = 596–911), measured during the first trimester; and MoBa ( N  = 39,243–57,170), reported by the mother at 15 weeks of gestation. We used multivariable regression to estimate the association of paternal BMI with the outcomes of interest adjusting (where available) for paternal age, number of children, education, smoking, and alcohol intake around the time of their partners’ pregnancy, as well as their partners’ BMI to account for the correlation between maternal and paternal BMI due to assortative mating or shared lifestyle [ 74 , 75 ] (correlation coefficients ranging from 0.17 in ALSPAC to 0.24 in MoBa). Results were then contrasted between the mutually adjusted maternal and paternal BMI (negative control) analyses. The adjusted maternal regression estimates used for comparison with paternal BMI associations in the negative control analysis differ from the multivariable regression estimates used in the main analysis (that are compared to the Mendelian randomisation estimates). In the paternal negative control comparison, the maternal regression estimates were additionally adjusted for paternal BMI and paternal confounders and therefore restricted to studies reporting both maternal and paternal BMI. Similar estimates between maternal and paternal BMI analyses indicate maternal BMI is unlikely to be a cause of pregnancy and perinatal outcomes via intrauterine mechanisms assuming comparable sources of biases. Conversely, associations that are specific or stronger in the maternal compared to the paternal BMI analyses would support a causal effect of maternal BMI.

Patient and public involvement

The current research was not informed by patient and public involvement because it used secondary data. This means that patients and the public were not involved in setting the research question or the outcome measures, nor were they involved in developing plans for the design or implementation of the study. No study participants were asked to advise on interpretation or writing up of results. The results of the research will be disseminated to study participants on request, and to stakeholders and the broader public as relevant.

Study and participant characteristics

The characteristics of the 14 included studies are shown in Table 1 . Mean maternal BMI ranged from 23.0 to 28.5 kg/m 2 across studies, and mean maternal age ranged from 25 to 31 years old. The maximum sample size from each study ranged from 356 (NFBC1966) to 190,879 (FinnGen). The number of cases ranged from 107 for miscarriage in the index pregnancy (used in multivariable regression and paternal negative control analyses) to 78,472 for breastfeeding initiation (Table 2 ).

Main analyses results

Results for the main multivariable regression (fully adjusted model) and Mendelian randomisation (IVW) analyses are shown in Figs. 3 and 4 (binary outcomes) and Additional file 3 : Supplementary Fig. 1 (continuous outcomes).

Comparison of A adjusted multivariable regression and main Mendelian randomisation estimates and B mutually adjusted multivariable regression estimates and paternal negative control (exposure, paternal body mass index)—for the association of maternal body mass index with binary outcomes (Part 1). Paternal BMI was used as a negative control exposure to explore the potential presence, direction, and magnitude of bias in multivariable estimates for associations of maternal BMI with outcomes.. Results are expressed as odds ratios per SD unit of maternal BMI and paternal BMI for ‘Multivariable regression’ and ‘Paternal negative control’, respectively. Multivariable regression results were adjusted for paternal BMI, maternal age, parity, education, smoking during pregnancy, alcohol use during pregnancy, and offspring sex where available. Paternal negative control results were adjusted for maternal BMI, paternal age, number of children (ALSPAC only), paternal education, paternal smoking, paternal alcohol use, and offspring sex. BMI, body mass index; NICU, neonatal intensive care unit

Comparison of A adjusted multivariable regression and main Mendelian randomisation estimates and B mutually adjusted multivariable regression estimates and paternal negative control (exposure, paternal body mass index)—for the association of maternal body mass index with binary outcomes (Part 2). Paternal BMI was used as a negative control exposure to explore the potential presence, direction, and magnitude of bias in multivariable estimates for associations of maternal BMI with outcomes.. Results are expressed as odds ratios per SD unit of maternal BMI and paternal BMI for ‘Multivariable regression’ and ‘Paternal negative control’, respectively. Multivariable regression results were adjusted for paternal BMI, maternal age, parity, education, smoking during pregnancy, alcohol use during pregnancy, and offspring sex where available. Paternal negative control results were adjusted for maternal BMI, paternal age, number of children (ALSPAC only), paternal education, paternal smoking, paternal alcohol use, and offspring sex. BMI, body mass index; NICU, neonatal intensive care unit

In the main multivariable regression analyses, maternal BMI was associated with 19 out of the 20 binary outcomes. Higher maternal BMI was associated with a higher risk of miscarriage, stillbirth, HDP, GH, PE, GDM, pre-labour membrane rupture, induction of labour, caesarean section, PTB, LGA, high birthweight, low Apgar score at 1 min, low Apgar score at 5 min, and NICU admission. In addition, women with higher BMI were less likely to have maternal anaemia, have a baby SGA or with low birthweight, and initiate breastfeeding (Figs. 3 and 4 ). There was little evidence of maternal BMI being associated with the risk of perinatal depression (Fig.  3 ). Higher maternal BMI was associated with higher values of most continuous outcomes (i.e. birthweight, birth length, and ponderal index) (Additional file 3 : Supplementary Fig. 1).

For the Mendelian randomisation analyses, we estimated that the total R 2 and mean F -statistic for the association of SNPs with BMI were 2.7% and 36, respectively, for the set of 97 SNPs using female-specific data from the GIANT GWAS. We observed a positive correlation ( r  = 0.67) between SNP-BMI estimates from females in the GIANT GWAS and SNP-BMI (pre-/early-pregnancy) estimates pooled across participating cohorts (Additional file 3 : Supplementary Fig. 2). In agreement with multivariable regression analyses, findings from Mendelian randomisation indicated that higher maternal BMI is related to higher risk of HDP, GH, PE, GDM, pre-labour membrane rupture, induction of labour, caesarean section, LGA, high birthweight, low Apgar score at 1 min, NICU admission, lower risk of having maternal anaemia, a SGA baby, lower odds of initiating breastfeeding, and not associated with perinatal depression. On the other hand, in contrast with multivariable regression analyses, Mendelian randomisation findings did not provide support for a positive association of maternal BMI with miscarriage, stillbirth, and PTB. As expected, given the lower statistical power, confidence intervals were wider for Mendelian randomisation compared to multivariable regression analyses and included the null value for some of these outcomes (Figs. 3 and 4 ). For two binary outcomes (i.e. low Apgar score at 5 min and low birthweight), it was less clear whether estimates from multivariable and Mendelian randomisation are in agreement given the substantial uncertainty in the latter. For most continuous outcomes (i.e. birthweight, birth length, and ponderal index), findings from Mendelian randomisation indicated that higher maternal BMI was associated with higher values of continuous outcomes in agreement with multivariable regression analyses (Additional file 3 : Supplementary Fig. 1).

Paternal negative control results supported the role of maternal BMI on stillbirth, HDP, GH, PE, GDM, maternal anaemia, pre-labour membrane rupture, induction of labour, caesarean section, SGA, LGA, high birthweight, low Apgar score at 1 min, NICU admission, and breastfeeding initiation (Figs. 3 and 4 ). The association of paternal BMI with maternal perinatal depression was also close to the null, consistent with maternal multivariable and Mendelian randomisation results. Associations with miscarriage, PTB, low birthweight, and low Apgar score at 5 min were imprecise and/or more similar in direction and magnitude between paternal and maternal BMI analyses. Results for continuous outcomes were strongly attenuated for paternal BMI in relation to birthweight and length (Additional file 3 : Supplementary Fig. 3).

Sensitivity analyses

Overall, findings from the main multivariable regression analyses were consistent across studies (Additional file 3 : Supplementary Fig. 4), when using random-effect metanalyses (Additional file 3 : Supplementary Fig. 5), and with minimally adjusted models (Additional file 3 : Supplementary Fig. 6). Between-study heterogeneity was substantial (i.e. Cochrane’s Q p -value < 0.05) for GDM, maternal anaemia, low Apgar score at 1 min, gestational age, and birthweight (Additional file 3 : Supplementary table 4).

Overall, findings from the main Mendelian randomisation analyses were not driven by any individual study as indicated by the leave-one-out analyses, although in some cases removing one study resulted in attenuation and substantial imprecision of effect estimates, such as for GDM when removing FinnGen and for delivery outcomes when removing MoBa (Additional file 3 : Supplementary Fig. 7). Results were similar when using fixed- or random-effect meta-analyses to pool SNP-outcome estimates across studies (Additional file 3 : Supplementary Fig. 8). There was evidence of substantial SNP heterogeneity in the IVW analyses of maternal BMI with 11 out of 20 binary outcomes and 1 out of 4 continuous outcomes (Additional file 2 : Supplementary table 5). Despite that, there was no clear evidence of directional pleiotropy as evidenced by the MR-Egger intercept test (except for GDM and gestational age) (Additional file 2: Supplementary table 5). Furthermore, Mendelian randomisation results were generally consistent when using different Mendelian randomisation methods (Additional file 3: Supplementary Fig. 9), although estimates from MR-Egger were imprecise for some outcomes. Effect estimates adjusting for offspring genotype were more imprecise due to the smaller sample size; however, overall, point estimates were not substantially different compared to the main analyses with a few exceptions, such as pre-labour rupture of membranes, LGA, and high birthweight, where adjusted results were attenuated (Additional file 3 : Supplementary Fig. 10).

Findings from the main paternal negative control analyses were consistent between studies (Additional file 3 : Supplementary Fig. 11 for maternal associations additionally adjusted for partners BMI and Additional file 3 : Supplementary Fig. 12 for paternal associations) and when comparing different models (Additional file 3 : Supplementary Figs. 13–15). Findings from the main multivariable regression analyses were similar when stratified by BMI taken pre-pregnancy compared to during pregnancy (Additional file 3 : Supplementary Fig. 16).

By triangulating different analytical approaches, our findings are compatible with higher maternal BMI being causally related to 14 out of 20 pregnancy and perinatal outcomes, including a higher risk of HDP, GH, PE, GDM, pre-labour membrane rupture, induction of labour, caesarean section, LGA, high birthweight, low Apgar score at 1 min, NICU admission, and lower odds of maternal anaemia, SGA, or breastfeeding initiation. In addition, we did not find supportive evidence for a relation of maternal BMI with perinatal depression. For other outcomes, evidence is uncertain due to inconsistencies across multiple approaches (i.e. multivariable regression results for miscarriage, stillbirth, and PTB were not supported by Mendelian randomisation) or substantial imprecision in effect estimates from Mendelian randomisation (i.e. low birthweight and low Apgar score at 5 min).

Consistent with our results, a previous study using multivariable regression reported higher maternal BMI (across the whole distribution) was associated with increased risk of HDP, GDM and LGA, and reduced risk of SGA based on data from 265,270 mother–offspring pairs (samples partly overlapping with our study) [ 10 ]. In addition, there was some evidence of a non-linear association with odds of PTB, which were higher in women who were underweight or obese [ 10 ]. In agreement with these findings, a larger study (9,282,486 mother–infant pairs in the USA) focussed on offspring outcomes indicated that higher maternal BMI was associated with a higher risk of high birthweight, LGA, and low Apgar score and reported a non-linear relationship with PTB risk [ 76 ]. Other observational studies using multivariable regression have reported that maternal BMI is associated with a higher risk of stillbirths [ 77 ], induction [ 78 ], caesarean section [ 78 ], and not initiating breastfeeding [ 79 ]. Previous Mendelian randomisation studies have focused on a limited set of outcomes and are supportive of higher maternal BMI being related to higher mean offspring birthweight [ 4 , 27 , 80 ] ( N  ~ 9,000 to 400,000) and GDM [ 81 ] ( N  = 5485 cases and 347,856 controls).

Recent systematic reviews of randomised controlled trials (RCTs) of diet and physical activity during pregnancy ( N range: 12,526–34,546) reported some evidence of reduced risk of GDM, LGA, and caesarean section in those randomised to the intervention, but no effect or mixed results of the intervention on HDP, PTB, and NICU admission [ 82 , 83 , 84 ]. Of note, these studies aimed at managing weight gain during pregnancy rather than targeting weight reduction prior to pregnancy with a modest mean difference of − 0.7 to − 1.2 kg between women in the intervention compared to those randomised to standard care. In addition, evidence for many outcomes is uncertain due to the relatively small number of cases.

Although mechanisms are not fully understood, higher maternal BMI is likely to influence a range of processes that are involved in the aetiology of some of the outcomes of interest, such as insulin resistance, endothelial dysfunction, inflammation, and susceptibility to infection [ 85 ]. In addition, maternal dysmetabolism resulting from excess adiposity has a well-recognised impact on maternal circulating nutrients, such as glucose, lipids, and amino acids, some of which can cross the placenta and influence offspring outcomes, such as growth [ 4 , 86 , 87 ].

Strengths and limitations

Key strengths of this study include exploring the potential role of maternal BMI on a wide range of pregnancy and perinatal outcomes in large samples from multiple studies using different approaches. The credibility of findings from each approach relies on the plausibility of assumptions that are often not possible to verify, such as no unmeasured confounding in multivariable regression, similar confounding, selection and measurement error between paternal and maternal BMI analyses, and no confounding or horizontal pleiotropy in Mendelian randomisation. Therefore, results in agreement across approaches strengthen the evidence on the relation of maternal BMI with the outcome. Where possible, we explored the plausibility of assumptions underlying each method. In particular, we conducted extensive sensitivity analyses to explore the plausibility of the core Mendelian randomisation assumptions and found overall these did not suggest Mendelian randomisation results were driven by weak, invalid instruments or confounding by offspring genotype.

Key limitations of this study are as follows. First, despite the large scale of our study, statistical power varied across outcomes as some outcomes have lower prevalence and/or were not collected in all cohorts. Second, despite our efforts to capture the best and most homogeneous definition for outcomes across studies, this was not always possible as exemplified by GDM, for which the data collected was notably variable across studies (e.g. from self-report to medical records-derived information), and index miscarriage (which was used for multivariable regression and paternal negative control analyses but is poorly captured in birth cohorts during the early pregnancy period). Third, while we were interested in maternal pre-pregnancy BMI, only maternal weight reflecting early-/mid-pregnancy was available in four studies. Fourth, our analyses assumed a linear effects of BMI, which may not be the case for some outcomes like PTB, and were restricted to women of European ancestry given most studies had scarce data on women from other ancestries. While this reduces the risk of confounding by ethnicity or population structure, it may limit the generalisability to other populations of pregnant women.

Our findings support a causal role for higher maternal BMI on a range of adverse pregnancy and perinatal outcomes. Given the high prevalence of overweight and obesity, our findings emphasise the need for development and testing of pre-conception interventions to support women maintaining a healthy BMI. This should be a key target to reduce the burden of obstetric and neonatal complications.

Availability of data and materials

In order to protect participant confidentiality, supporting data cannot be made openly available. Bona fide researchers can apply for access to study-specific executive committees. Summary association data for FinnGen is publicly available at https://www.finngen.fi/en/access_results . Researchers can apply for access to the UK Biobank data via the Access Management System (AMS) ( https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access ).

Global Burden of Disease Collaborative Network: Global Burden of Disease Study 2019 (GBD 2019) results. In . Seattle, USA: Institute for Health Metrics and Evaluation (IHME); 2021.

Marchi J, Berg M, Dencker A, Olander E, Begley C. Risks associated with obesity in pregnancy, for the mother and baby: a systematic review of reviews. Obes Rev. 2015;16(8):621–38.

Article   CAS   PubMed   Google Scholar  

Aune D, Saugstad OD, Henriksen T, Tonstad S. Maternal body mass index and the risk of fetal death, stillbirth, and infant death: a systematic review and meta-analysis. JAMA. 2014;311(15):1536–46.

Tyrrell J, Richmond RC, Palmer TM, Feenstra B, Rangarajan J, Metrustry S, Cavadino A, Paternoster L, Armstrong LL, De Silva NMG. Genetic evidence for causal relationships between maternal obesity-related traits and birth weight. JAMA. 2016;315(11):1129–40.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Poobalan AS, Aucott LS, Gurung T, Smith WCS, Bhattacharya S. Obesity as an independent risk factor for elective and emergency caesarean delivery in nulliparous women–systematic review and meta-analysis of cohort studies. Obes Rev. 2009;10(1):28–35.

Heslehurst N, Simpson H, Ells L, Rankin J, Wilkinson J, Lang R, Brown T, Summerbell C. The impact of maternal BMI status on pregnancy outcomes with immediate short-term obstetric resource implications: a meta-analysis. Obes Rev. 2008;9(6):635–83.

Molyneaux E, Poston L, Ashurst-Williams S, Howard LM. Obesity and mental disorders during pregnancy and postpartum: a systematic review and meta-analysis. Obstet Gynecol. 2014;123(4):857.

Article   PubMed   PubMed Central   Google Scholar  

Molyneaux E, Pasupathy D, Kenny L, McCowan L, North R, Dekker G, Walker J, Baker PN, Poston L, Howard L. Socio-economic status influences the relationship between obesity and antenatal depression: data from a prospective cohort study. J Affect Disord. 2016;202:124–7.

Turcksin R, Bel S, Galjaard S, Devlieger R. Nutrition c: Maternal obesity and breastfeeding intention, initiation, intensity and duration: a systematic review. Matern Child Nutr. 2014;10(2):166–83.

Article   PubMed   Google Scholar  

Santos S, Voerman E, Amiano P, Barros H, Beilin LJ, Bergstrom A, Charles MA, Chatzi L, Chevrier C, Chrousos GP, et al. Impact of maternal body mass index and gestational weight gain on pregnancy complications: an individual participant data meta-analysis of European North American and Australian cohorts. BJOG. 2019;126(8):984–95.

Lutsiv O, Mah J, Beyene J, McDonald SD. The effects of morbid obesity on maternal and neonatal health outcomes: a systematic review and meta-analyses. Obes Rev. 2015;16(7):531–46.

Dachew BA, Ayano G, Betts K, Alati R. The impact of pre-pregnancy BMI on maternal depressive and anxiety symptoms during pregnancy and the postpartum period: A systematic review and meta-analysis. J Affect Disord. 2021;281:321–30.

Lawlor DA, Tilling K, Davey Smith G. Triangulation in aetiological epidemiology. Int J Epidemiol. 2016;45(6):1866–86.

PubMed   Google Scholar  

Munafo MR, Davey Smith G. Robust research needs many lines of evidence. Nature. 2018;553(7689):399–401.

Lawlor DA, Harbord RM, Sterne JA, Timpson N, Davey Smith G. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat Med. 2008;27(8):1133–63.

Smith GD, Lawlor DA, Harbord R, Timpson N, Day I, Ebrahim S. Clustered environments and randomized genes: a fundamental distinction between conventional and genetic epidemiology. PLoS Med. 2007;4(12): e352.

Lipsitch M, Tchetgen Tchetgen E, Cohen T. Negative controls: a tool for detecting confounding and bias in observational studies. Epidemiology. 2010;21(3):383–8.

Sanderson E, Macdonald-Wallis C, Davey Smith G. Negative control exposure studies in the presence of measurement error: implications for attempted effect estimate calibration. Int J Epidemiol. 2018;47(2):587–96.

Ebrahim S, Davey Smith G. Mendelian randomization: can genetic epidemiology help redress the failures of observational epidemiology? Hum Genet. 2008;123(1):15–33.

Smith GD, Ebrahim S. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol. 2003;32(1):1–22.

Sanderson E, Glymour MM, Holmes MV, Kang H, Morrison J, Munafò MR, Palmer T, Schooling CM, Wallace C, Zhao Q, et al. Mendelian randomization Nature Reviews Methods Primers. 2022;2(1):6.

Yang Q, Borges MC, Sanderson E, Magnus MC, Kilpi F, Collings PJ, Soares AL, West J, Magnus P, Wright J, et al. Associations between insomnia and pregnancy and perinatal outcomes: Evidence from mendelian randomization and multivariable regression analyses. PLoS Med. 2022;19(9): e1004090.

Boyd A, Golding J, Macleod J, Lawlor DA, Fraser A, Henderson J, Molloy L, Ness A, Ring S, Davey Smith G. Cohort Profile: the ’children of the 90s’–the index offspring of the Avon Longitudinal Study of Parents and Children. Int J Epidemiol. 2013;42(1):111–27.

Fraser A, Macdonald-Wallis C, Tilling K, Boyd A, Golding J, Davey Smith G, Henderson J, Macleod J, Molloy L, Ness A, et al. Cohort Profile: the Avon Longitudinal Study of Parents and Children: ALSPAC mothers cohort. Int J Epidemiol. 2013;42(1):97–110.

Taylor K, McBride N, Goulding N, Burrows K, Mason D, Pembrey L, Yang T, Azad R, Wright J, Lawlor D: Metabolomics datasets in the Born in Bradford cohort [version 2; peer review: 1 approved, 1 approved with reservations]. Wellcome Open Res 2021, 5(264).

Wright J, Small N, Raynor P, Tuffnell D, Bhopal R, Cameron N, Fairley L, Lawlor DA, Parslow R, Petherick ES, et al. Cohort Profile: the Born in Bradford multi-ethnic family cohort study. Int J Epidemiol. 2013;42(4):978–91.

Bond TA, Richmond RC, Karhunen V, Cuellar-Partida G, Borges MC, Zuber V, Couto Alves A, Mason D, Yang TC, Gunter MJ, et al. Exploring the causal effect of maternal pregnancy adiposity on offspring adiposity: Mendelian randomisation using polygenic risk scores. BMC Med. 2022;20(1):34.

Nohr EA, Timpson NJ, Andersen CS, Davey Smith G, Olsen J, Sorensen TI. Severe obesity in young women and reproductive health: the Danish National Birth Cohort. PLoS ONE. 2009;4(12): e8444.

Paternoster L, Evans DM, Nohr EA, Holst C, Gaborieau V, Brennan P, Gjesing AP, Grarup N, Witte DR, Jorgensen T, et al. Genome-wide population-based association study of extremely overweight young adults–the GOYA study. PLoS ONE. 2011;6(9): e24303.

Schnurr TM, Morgen CS, Borisevich D, Beaumont RN, Engelbrechtsen L, Angquist L, Have CT, Freathy RM, Smith GD, Nohr EA, et al. The influence of transmitted and non-transmitted parental BMI-associated alleles on the risk of overweight in childhood. Sci Rep. 2020;10(1):4806.

Bliddal M, Broe A, Pottegard A, Olsen J, Langhoff-Roos J. The danish medical birth register. Eur J Epidemiol. 2018;33(1):27–36.

Ryckman KK, Feenstra B, Shaffer JR, Bream EN, Geller F, Feingold E, Weeks DE, Gadow E, Cosentino V, Saleme C, et al. Replication of a genome-wide association study of birth weight in preterm neonates. J Pediatr. 2012;160(1):19–24 e14.

Olsen J, Melbye M, Olsen SF, Sorensen TI, Aaby P, Andersen AM, Taxbol D, Hansen KD, Juhl M, Schow TB, et al. The Danish National Birth Cohort–its background, structure and aim. Scand J Public Health. 2001;29(4):300–7.

Knight B, Shields BM, Hattersley AT. The Exeter Family Study of Childhood Health (EFSOCH): study protocol and methodology. Paediatr Perinat Epidemiol. 2006;20(2):172–9.

Manichaikul A, Mychaleckyj JC, Rich SS, Daly K, Sale M. Chen W-MJB: Robust relationship inference in genome-wide association studies. 2010;26(22):2867–73.

CAS   Google Scholar  

Abraham G. Inouye MJPo: Fast principal component analysis of large-scale genome-wide data. 2014;9(4): e93766.

Google Scholar  

Kurki MI, Karjalainen J, Palta P, Sipilä TP, Kristiansson K, Donner K, Reeve MP, Laivuori H, Aavikko M, Kaunisto MA et al : FinnGen: Unique genetic insights from combining isolated population and national health register data. medRxiv 2022:2022.2003.2003.22271360.

Kiiskinen T, Mars NJ, Palviainen T, Koskela J, Ramo JT, Ripatti P, Ruotsalainen S, FinnGen GC, Palotie A, Madden PAF, et al. Genomic prediction of alcohol-related morbidity and mortality. Transl Psychiatry. 2020;10(1):23.

Magee LA, Pels A, Helewa M, Rey E, von Dadelszen P. Canadian Hypertensive Disorders of Pregnancy Working G: Diagnosis, evaluation, and management of the hypertensive disorders of pregnancy: executive summary. J Obstet Gynaecol Can. 2014;36(5):416–41.

Fenton TR, Kim JH. A systematic review and meta-analysis to revise the Fenton growth chart for preterm infants. BMC Pediatr. 2013;13:59.

Kooijman MN, Kruithof CJ, van Duijn CM, Duijts L, Franco OH. van IMH, de Jongste JC, Klaver CC, van der Lugt A, Mackenbach JP et al : The Generation R Study: design and cohort update 2017. Eur J Epidemiol. 2016;31(12):1243–64.

Coolman M, de Groot CJ, Jaddoe VW, Hofman A, Raat H, Steegers EA. Medical record validation of maternally reported history of preeclampsia. J Clin Epidemiol. 2010;63(8):932–7.

Group HSCR. The Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study. Int J Gynaecol Obstet. 2002;78(1):69–77.

Article   Google Scholar  

Group HSCR, Metzger BE, Lowe LP, Dyer AR, Trimble ER, Chaovarindr U, Coustan DR, Hadden DR, McCance DR, Hod M, et al. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med. 2008;358(19):1991–2002.

Laurie CC, Doheny KF, Mirel DB, Pugh EW, Bierut LJ, Bhangale T, Boehm F, Caporaso NE, Cornelis MC, Edenberg HJ, et al. Quality control and quality assurance in genotypic data for genome-wide association studies. Genet Epidemiol. 2010;34(6):591–602.

Guxens M, Ballester F, Espada M, Fernandez MF, Grimalt JO, Ibarluzea J, Olea N, Rebagliato M, Tardon A, Torrent M, et al. Cohort Profile: the INMA–INfancia y Medio Ambiente–(Environment and Childhood) Project. Int J Epidemiol. 2012;41(4):930–40.

Magnus P, Birke C, Vejrup K, Haugan A, Alsaker E, Daltveit AK, Handal M, Haugen M, Høiseth G, Knudsen GP, et al. Cohort profile update: The Norwegian mother and child cohort study (MoBa). Int J Epidemiol. 2016;45(2):382–8.

Rantakallio P: Groups at risk in low birth weight infants and perinatal mortality. Acta Paediatr Scand 1969, 193:Suppl 193:191+.

Nordström T, Miettunen J, Auvinen J, Ala-Mursula L, Keinänen-Kiukaanniemi S, Veijola J, Järvelin M-R, Sebert S, Männikkö M. Cohort Profile: 46 years of follow-up of the Northern Finland Birth Cohort 1966 (NFBC1966). Int J Epidemiol. 2021;50(6):1786–1787j.

Sabatti C, Service SK, Hartikainen A-L, Pouta A, Ripatti S, Brodsky J, Jones CG, Zaitlen NA, Varilo T, Kaakinen M et al : Genome-wide association analysis of metabolic traits in a birth cohort from a founder population. Nature Genetics 2009, 41(1):35–46.

Järvelin MR, Hartikainen-Sorri AL, Rantakallio P. Labour induction policy in hospitals of different levels of specialisation. Br J Obstet Gynaecol. 1993;100(4):310–5.

Taanila A, Ebeling H, Kotimaa A, Moilanen I, Järvelin MR. Is a large family a protective factor against behavioural and emotional problems at the age of 8 years? Acta Paediatr. 2004;93(4):508–17.

Mitchell R, Hemani, G., Dudding, T., Corbin, L., Harrison, S., Paternoster, L: UK Biobank Genetic Data: MRC-IEU Quality Control, version 2 - Datasets - data.bris. data.bris. 2018.

Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, Motyer A, Vukcevic D, Delaneau O, O’Connell J, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203–9.

O’Connell J, Sharp K, Shrine N, Wain L, Hall I, Tobin M, Zagury JF, Delaneau O, Marchini J. Haplotype estimation for biobank-scale data sets. Nat Genet. 2016;48(7):817–20.

Guillemette L, Allard C, Lacroix M, Patenaude J, Battista M-C, Doyon M, Moreau J, Ménard J, Bouchard L. Ardilouze J-L: Genetics of Glucose regulation in Gestation and Growth (Gen3G): a prospective prebirth cohort of mother–child pairs in Sherbrooke Canada. BMJ Open. 2016;6(2): e010031.

Magnus P, Irgens LM, Haug K, Nystad W, Skjaerven R, Stoltenberg C, MoBa Study G: Cohort profile: the Norwegian Mother and Child Cohort Study (MoBa). Int J Epidemiol 2006; 35(5):1146-1150

Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, Downey P, Elliott P, Green J. Landray MJPm: UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. 2015;12(3):e1001779.

StataCorp: Stata Statistical Software: Release 15. In . : College Station, TX: StataCorp LLC; 2017.

team Rc: R: A Language and Environment ofr Statistical Computing. In . Edited by Computing RFfS. Vienna, Austria; 2017.

Lawlor DA. Commentary: Two-sample Mendelian randomization: opportunities and challenges. Int J Epidemiol. 2016;45(3):908–15.

Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, Powell C, Vedantam S, Buchkovich ML, Yang J, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518(7538):197–206.

Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson G, Jackson AU, Allen HL, Lindgren CM, Luan Ja, Mägi RJNg: Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. 2010, 42(11):937.

Yengo L, Sidorenko J, Kemper KE, Zheng Z, Wood AR, Weedon MN, Frayling TM, Hirschhorn J, Yang J, Visscher PM, et al. Meta-analysis of genome-wide association studies for height and body mass index in approximately 700000 individuals of European ancestry. Hum Mol Genet. 2018;27(20):3641–9.

Burgess S, Davies NM, Thompson SG. Bias due to participant overlap in two-sample Mendelian randomization. Genet Epidemiol. 2016;40(7):597–608.

Burgess S, Thompson SG. Bias in causal estimates from Mendelian randomization studies with weak instruments. Stat Med. 2011;30(11):1312–23.

Burgess S, Butterworth A, Thompson SG: Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol. 2013, 37(7):658-665

Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512–25.

Bowden J, Davey Smith G, Haycock PC. Burgess SJGe: Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator. 2016;40(4):304–14.

Hartwig FP, Davey Smith G, Bowden J: Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. 2017, 46(6):1985-1998.

Warrington NM, Beaumont RN, Horikoshi M, Day FR, Helgeland O, Laurin C, Bacelis J, Peng S, Hao K, Feenstra B, et al. Maternal and fetal genetic effects on birth weight and their relevance to cardio-metabolic risk factors. Nat Genet. 2019;51(5):804–14.

Lawlor D, Richmond R, Warrington N, McMahon G, Davey Smith G, Bowden J, Evans DM. Using Mendelian randomization to determine causal effects of maternal pregnancy (intrauterine) exposures on offspring outcomes: Sources of bias and methods for assessing them. Wellcome Open Res. 2017;2:11.

Evans DM, Moen GH, Hwang LD, Lawlor DA, Warrington NM. Elucidating the role of maternal environmental exposures on offspring health and disease using two-sample Mendelian randomization. Int J Epidemiol. 2019;48(3):861–75.

Brand JS, Gaillard R, West J, McEachan RRC, Wright J, Voerman E, Felix JF, Tilling K, Lawlor DA. Associations of maternal quitting, reducing, and continuing smoking during pregnancy with longitudinal fetal growth: Findings from Mendelian randomization and parental negative control studies. PLoS Med. 2019;16(11): e1002972.

Madley-Dowd P, Rai D, Zammit S, Heron J. Simulations and directed acyclic graphs explained why assortative mating biases the prenatal negative control design. J Clin Epidemiol. 2020;118:9–17.

Zong X, Wang H, Yang L, Guo Y, Zhao M, Magnussen CG, Xi B. Maternal pre-pregnancy body mass index categories and infant birth outcomes: A population-based study of 9 Million mother-infant pairs. Front Nutr. 2022;9: 789833.

Ellis JA, Brown CM, Barger B, Carlson NS. Influence of maternal obesity on labor induction: a systematic review and meta-analysis. J Midwifery Womens Health. 2019;64(1):55–67.

Ramji N, Quinlan J, Murphy P, Crane JM. The impact of maternal obesity on breastfeeding. J Obstet Gynaecol Can. 2016;38(8):703–11.

Thompson W, Beaumont R, Kuang A, Warrington N, Ji Y, Tyrrell J, Wood A, Scholtens D, Knight B, Evans D et al : Higher maternal adiposity reduces offspring birth weight if associated with a metabolically favourable profile. medRxiv 2020:2020.2005.2025.20112441.

Pervjakova N, Moen GH, Borges MC, Ferreira T, Cook JP, Allard C, Beaumont RN, Canouil M, Hatem G, Heiskala A et al : Multi-ancestry genome-wide association study of gestational diabetes mellitus highlights genetic links with type 2 diabetes. Hum Mol Genet 2022.

International Weight Management in Pregnancy Collaborative G: Effect of diet and physical activity based interventions in pregnancy on gestational weight gain and pregnancy outcomes: meta-analysis of individual participant data from randomised trials. BMJ 2017, 358:j3119.

Cantor AG, Jungbauer RM, McDonagh M, Blazina I, Marshall NE, Weeks C, Fu R, LeBlanc ES, Chou R. Counseling and behavioral interventions for healthy weight and weight gain in pregnancy: evidence report and systematic review for the US Preventive Services Task Force. JAMA. 2021;325(20):2094–109.

Teede HJ, Bailey C, Moran LJ, Bahri Khomami M, Enticott J, Ranasinha S, Rogozinska E, Skouteris H, Boyle JA, Thangaratinam S, et al. Association of antenatal diet and physical activity-based interventions with gestational weight gain and pregnancy outcomes: a systematic review and meta-analysis. JAMA Intern Med. 2022;182(2):106–14.

Catalano PM, Shankar K. Obesity and pregnancy: mechanisms of short term and long term adverse consequences for mother and child. BMJ. 2017;356: j1.

Barry C-JS, Lawlor DA, Shapland CY, Sanderson E, Borges MC: Using Mendelian randomisation to prioritise candidate maternal metabolic traits influencing offspring birthweight. Metabolites 2022, 12(6):537.

Zhao J, Stewart ID, Baird D, Mason D, Wright J, Zheng J, Gaunt TR, Evans DM, Freathy RM, Langenberg C et al : Causal effects of maternal circulating amino acids on offspring birthweight: a Mendelian randomisation study. medRxiv 2022:2022.2004.2015.22273911.

Download references

Acknowledgements

We acknowledge Ville Karhunen for his support in advising data analysis from NFBC1966 and NFBC1986. Details on study-specific acknowledgements are provided in Supplementary material (Additional file 1 : Supplementary methods).

The views expressed in this paper are those of the authors and do not necessarily reflect the views of any funders, person, or group listed in funding or acknowledgement statements.

This study was supported by the MRC Integrative Epidemiology Unit at the University of Bristol (MC_UU_00032/05), British Heart Foundation (AA/18/1/34219), the European Research Council under the European Union’s Seventh Framework Programme (FP/2007–2013)/ERC Grant Agreement (Grant number 669545), the European Union’s Horizon 2020 research and innovation programme under grant agreement No 733206 (LifeCycle), the US National Institutes of Health (R01 DK10324, U01 HG004415), the Bristol NIHR Biomedical Research Centre, and the Research Council of Norway through its Centres of Excellence funding scheme (project number 262700), and the Wellcome Trust [Grant number WT220390]. For the purpose of open access, the authors have applied a CC BY public copyright licence to any Author-Accepted Manuscript version arising from this submission.

MCB has received support from MRC Skills Development Fellowship (MR/P014054/1) and the University of Bristol Vice-Chancellor’s Fellowship. DAL is a British Heart Foundation Chair (CH/F/20/90003) and NIHR Senior Investigator (NF-0616–10102). JT is supported by an Academy of Medical Sciences (AMS) Springboard Award, which is supported by the AMS, the Wellcome Trust, GCRF, the Government Department of Business, Energy and Industrial Strategy, the British Heart Foundation and Diabetes UK [SBF004\1079]. RMF was funded by a Wellcome Trust and Royal Society Sir Henry Dale Fellowship (WT104150) and is now funded by a Wellcome Trust Senior Research Fellowship (WT220390). RG received funding from the Dutch Heart Foundation (grant number 2017T013), the Dutch Diabetes Foundation (grant number 2017.81.002), and the Netherlands Organization for Health Research and Development (NWO, ZonMW, grant number 543003109). VWVJ received a Consolidator Grant from the European Research Council (ERC-2014-CoG-648916). XL received support from the Nordic Center of Excellence in Health-Related e-Sciences. LS reports funding from a Carlsberg Foundation postdoctoral fellowship (CF15-0899). BF was supported by an Oak Foundation Fellowship and by a grant from the Novo Nordisk Foundation (12,955). MCM has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement number 947684). TAB is supported by the Medical Research Council (MRC) (UK) (MR/K501281/1), the NHMRC (Australia) (GNT1183074 and GNT1157714), and the British Heart Foundation Accelerator Award at the University of Bristol (AA/18/1/34219) and works in/is affiliated with a unit that is supported by the UK Medical Research Council (MC_UU_00011/6). LB is a senior research scholar from the Fonds de la recherche du Québec-Santé (FRQ-S) and a member of the FRQ-S-funded Centre de recherche du CHUS. MFH was supported by an American Diabetes Association (ADA) Pathways Accelerator Award (1–15-ACE-26). SEH and MCM are partly funded by the Research Council of Norway (project no. 320656) and through its Centres of Excellence funding scheme (project No 262700). JW and RMc are supported by the National Institute for Health and Care Research under its Applied Research Collaboration, Yorkshire and Humber (NIHR200166).

Details on study-specific funding are provided in Supplementary material (Additional file 1 : Supplementary methods).

Author information

Maria Carolina Borges and Gemma Clayton are joint first authors.

Authors and Affiliations

MRC Integrative Epidemiology Unit at the University of Bristol, Oakfield House, Oakfield Grove, Bristol, BS8 2BN, UK

Maria Carolina Borges, Gemma L. Clayton, Rachel M. Freathy, Alba Fernández-Sanlés, Ana Gonçalves Soares, Fanny Kilpi, Qian Yang, Rebecca C. Richmond, Tom A. Bond & Deborah A. Lawlor

Population Health Science, Bristol Medical School, University of Bristol, Bristol, UK

Maria Carolina Borges, Gemma L. Clayton, Alba Fernández-Sanlés, Ana Gonçalves Soares, Fanny Kilpi, Qian Yang, Rebecca C. Richmond, Tom A. Bond & Deborah A. Lawlor

Institute of Biomedical and Clinical Science, College of Medicine and Health, University of Exeter, Exeter, UK

Rachel M. Freathy, Andrew T. Hattersley & Jessica Tyrrell

The Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands

Janine F. Felix, Romy Gaillard & Vincent W. V. Jaddoe

Department of Pediatrics, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands

Bradford Institute for Health Research, Bradford Teaching Hospitals NHS Trust, Bradford, UK

Rosemary R. C. McEachan, Jane West & John Wright

Department of Epidemiology Research, Statens Serum Institut, Copenhagen, Denmark

Xueping Liu, Line Skotte & Bjarke Feenstra

Department of Preventive Medicine and Public Health, University of the Basque Country, Leioa, Spain

Amaia Irizar

BIODONOSTIA Health Research Institute, Paseo Dr. Beguiristain, 20014, San Sebastian, Spain

CIBER Epidemiología y Salud Pública (CIBERESP), Madrid, Spain

Amaia Irizar, Mario Murcia, Mariona Bustamante & Martine Vrijheid

MRC Centre for Environment and Health, School of Public Health, Imperial College London, London, UK

Barbara Bodinier & Tom A. Bond

Department of Preventive Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL, USA

Denise M. Scholtens

Institute of Clinical Research, University of Southern Denmark, Odense, Denmark

Ellen A. Nohr

Department of Epidemiology and Biostatistics, Imperial College London, London, UK

Tom A. Bond

The University of Queensland Diamantina Institute, The University of Queensland, Brisbane, Australia

Division of Endocrinology, Metabolism, and Molecular Medicine, Department of Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL, USA

M. Geoffrey Hayes & William L. Lowe Jr

Department of Biochemistry and Functional Genomics, Faculty of Medicine and Health Sciences, Université de Sherbrooke, Sherbrooke, Québec, Canada

Luigi Bouchard

Epidemiology and Environmental Health Joint Research Unit, FISABIO-Universitat Jaume I-Universitat de València, Valencia, Spain

Mario Murcia

ISGlobal, Institute for Global Health, Barcelona, Spain

Mariona Bustamante & Martine Vrijheid

Universitat Pompeu Fabra (UPF), Barcelona, Spain

Faculty of Medicine, School of Public Health, Imperial College London, London, UK

Marc Chadeau-Hyam & Marjo-Riitta Jarvelin

Centre de Recherche du Centre Hospitalier Universitaire de Sherbrooke (CR-CHUS), Sherbrooke, Québec, Canada

Patrice Perron

Department of Medicine, Faculty of Medicine and Health Sciences, University of Sherbrooke, Sherbrooke, Québec, Canada

Centre for Fertility and Health, Norwegian Institute of Public Health, Oslo, Norway

Per Magnus, Siri E. Håberg & Maria Magnus

Department of Population Medicine, Harvard Medical School, Harvard Pilgrim Health Care Institute, Boston, MA, USA

Marie-France Hivert

Diabetes Unit, Massachusetts General Hospital, Boston, MA, USA

Department of Public Health, Section of Epidemiology, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark

Thorkild I. A. Sørensen

Novo Nordisk Foundation Center for Basic Metabolic Diseases, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark

Center For Life-Course Health Research, Faculty of Medicine, University of Oulu, Oulu, Finland

Sylvain Serbert

You can also search for this author in PubMed   Google Scholar

Contributions

DAL designed the study. MCB and GLC developed the analysis plan, performed data analyses, and wrote the first draft of the manuscript. RMF, JFF, AFS, AGS, FK, QY, RRCM, RCR, XL, LS, AI, ATH, BB, DMS, EAN, TAB, MGH, JW1, JT, JW2, LB, MM1, MB, MCH, MRJ, MV, PP, PM, RG, VWVJ, WLLJ, BF, MFH, TIAS, SHE, SS, and MM2 have made substantial contributions to acquiring data, analysing study-specific data, interpreting results, and revising the draft for important intellectual content. MCB, GLC, and DAL will act as guarantors of the study. All authors read and approved the final version of the manuscript.

Authors’ Twitter handles

Twitter handles: @MCarol_Borges (Maria Carolina Borges).

Twitter handles: @mrc_ieu (MRC Integrative Epidemiology Unit).

Twitter handles: @clayton_gem (Gemma Clayton).

Corresponding authors

Correspondence to Maria Carolina Borges or Deborah A. Lawlor .

Ethics declarations

Ethics approval and consent to participate.

ALSPAC: Ethical approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committees (NHS Haydock REC: 10/H1010/70).

BiB: Ethical approval for the study was granted by the Bradford National Health Service Research Ethics Committee (ref 06/Q1202/48).

DNBC-GOYA: The study was approved by the regional scientific ethics committee and by the Danish Data Protection Board.

DNBC-PTB: Ethical approval was obtained from the Regional Scientific Ethical Committee of Copenhagen and the study was also approved by the Danish Data Protection Agency.

EFSOCH: All women gave informed consent and ethical approval was obtained from the local review committee.

FinnGen: The Coordinating Ethics Committee of the Helsinki and Uusimaa Hospital District has approved the FinnGen consortium (Nr HUS/990/2017), and the ethical approval of each individual study has been described in detail elsewhere [35].

Gen3G: Ethics approval was obtained from the Centre Hospitalier Universitaire de Sherbrooke (CHUS) Ethics Review Board for Studies with Humans.

Generation R: The study protocol was approved by the Medical Ethical Committee of the Erasmus MC, University Medical Center Rotterdam and informed consent was obtained for all participants.

HAPO: The protocol was approved by the institutional review board at each field center. All participants gave written informed consent. An external data and safety monitoring committee provided oversight.

INMA: Informed consent was obtained from all participants and the study was approved by the Hospital Ethics Committees in each participating region.

MoBa: The establishment of MoBa and initial data collection was based on a licene from the Norwegian Data Protection Agency and approval from The Regional Committees for Medical and Health Research Ethics. The MoBa cohort is currently regulated by the Norwegian Health Registry Act. The current study was approved by The Regional Committees for Medical and Health Research Ethics of South/East Norway (ref 2018/1256).

NFBC1966 and 1986: An informed consent for the use of the data including DNA was obtained from all subjects. NFBC1966 received ethical approval from the Ethics Committee of Northern Ostrobothnia Hospital District (EETTMK: 94/2011) and Oulu University, Faculty of Medicine, Oulu, Finland. NFBC1986 received ethical approval from the Ethics Committee of Northern Ostrobothnia Hospital District (EETTMK: 108/ 2017) and Oulu University, Faculty of Medicine, Oulu, Finland.

UK Biobank: Ethical approval for UKB was obtained from the North West Multi-centre Research Ethics Committee (MREC), and our study was performed under UKB application number 23938.

Details of ethical approval and consent to participate of each study can also be found in Additional file 1 : Supplementary Methods.

Consent for publication

Not applicable.

Competing interests

DAL receives support from several national and international government and charitable research funders, as well as from Medtronic Ltd and Roche Diagnostics for research unrelated to that presented here. The other authors declare that they have no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1..

Supplementary methods. 

Additional file 2.

 Supplementary tables.

Additional file 3.

 Supplementary figures.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Borges, M.C., Clayton, G.L., Freathy, R.M. et al. Integrating multiple lines of evidence to assess the effects of maternal BMI on pregnancy and perinatal outcomes. BMC Med 22 , 32 (2024). https://doi.org/10.1186/s12916-023-03167-0

Download citation

Received : 29 March 2023

Accepted : 09 November 2023

Published : 29 January 2024

DOI : https://doi.org/10.1186/s12916-023-03167-0

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Body mass index
• Triangulation
• Mendelian randomisation

BMC Medicine

ISSN: 1741-7015

• Submission enquiries: [email protected]
• General enquiries: [email protected]

• Skip to main content
• Keyboard shortcuts for audio player

• Your Health
• Treatments & Tests
• Health Inc.
• Public Health

How to Thrive as You Age

Step aside bmi, body composition tests are on the rise. here's what to know.

Allison Aubrey

Body composition tests are more useful than BMI

Maria Fabrizio hide caption

The scale has never been a friend to Mana Mostatabi. Even back in high school, when she ran a quick 100m on the varsity track team, her BMI – a ratio of weight to height – put her in the overweight category.

“My dad always joked that I should be a wrestler,” Mostatabi says due to her build. Many professional athletes flunk BMI tests. Some are considered obese despite their fitness, and many doctors say it isn’t a helpful metric to focus on.

“BMI is a very crude measure,” says Dr. Richard Joseph , a physician at Brigham and Women’s Hospital who specializes in metabolic health. “It doesn’t tell me much about your underlying health,” he says. People can be a normal weight but have low muscle mass and high body fat, while others have higher body weight but are muscular and lean.

That’s why Mostatabi has found a new tool – a body composition scan – that measures her body fat and muscle mass, which are two key metrics of health. “It’s very affirming,” Mostatabi says. Over the last year, she has lost ten pounds of body fat and also gained several pounds of muscle. “This actually gives me information,” to track progress. “It really is motivating,” she says.

Body composition scans are becoming an increasingly popular way to gauge health and there are lots of different kinds.

An MRI ( magnetic resonance imaging ) is considered the gold standard, but it’s not practical for most people given the expense and access to medical imaging.

Dr. Joseph orders DEXA – dual-energy X-ray absorptiometry – scans for some of his patients. These scans measure bone mineral density, and also measure body composition and fat distribution. They typically cost more than $100, but prices have begun to drop in some areas as demand rises and more machines are available.

An option that’s taken off in gyms and workout studios, such as Anytime Fitness and Orangetheory Fitness is bioelectric impedance analysis, using devices such as the InBody or the Evolt 360 . Depending on the studio, the scans are often free with membership or are available for a small fee. This test is not as precise as an MRI or DEXA, but is reliable at tracking changes over time, as long as people follow directions.

Mana Mostatabi uses a body composition test to track her fitness. It's helped her lower her body fat and increase muscle. Mana Mostatabi hide caption

Mostatabi had her first InBody scan in January, when she signed up for a strength challenge at Orangetheory. “It’s a super simple process,” Mostatabi says, which takes less than a minute.

The device looks like a scale with two arms. “You step onto the machine,” which has a metal base, she explains and you hold onto the ends of the two arms which send a safe, low-level electrical current through the body, estimating fat and lean mass.

“How fast that current is returning to the electrodes gives a measurement of how much fat mass and muscle you have, because the current travels through those body tissues at different speeds,” explains Scott Brown, Vice President of fitness at Orangetheory. The technology is decades old , and has become increasingly popular with the advent of commercial devices and rising demand.

Mostatabi says you can’t feel anything during the test, and the results are sent directly to a smart phone app. Her first scan gave her a benchmark to improve upon.

She explains her fitness took a nosedive during the pandemic, and she was aiming for a fresh start this year. All winter she pushed herself through 60 minute classes that combined resistance training, weight lifting, cardio on a treadmill and rowing. “I was very diligent,” averaging about five to six classes per week.

Shots - Health News

Women who do strength training live longer. how much is enough.

“The InBody scan is the first tool I’ve ever used that shows gains,” she says. Mostatabi was accustomed to weighing herself, and recalls the frustration when the scale would not budge. Now, even if she hasn’t lost weight, she knows her body is healthier, with a lower percentage of body fat and an increase in muscle.

“It’s just incredibly empowering,” she says, knowing that women who weight train and build strength can expect to live longer and healthier lives.

On average, women gain less muscle at a slower rate compared to men. During Orangetheory’s eight week ‘transformation challenge’ the company found that males who were focused on muscle gains added about .83 lbs of muscle over eight weeks, on average, compared to a half pound of muscle gain for females. Though across all participants there was only about .1 lb of average muscle gain.

Eight weeks is not a lot of time to gain muscle, explains Brown. And he says it’s important to set “realistic goals and targets'' given the variability from person to person. The ability to build new muscle is influenced by gender, age and genetics.

Dr. Joseph says the reason it can be helpful to know your muscle mass is because studies show that strength is a predictor of longevity. Also, loss of muscle increases the risk of falling, which is a top cause of death from injury among older people. “A lot of people are under-muscled,” Joseph says.

When it comes to body fat, having too much can increase the risk of metabolic disease, especially visceral fat , which surrounds the abdominal organs including the stomach, liver and intestines. “It’s inflammatory,” Joseph says and drives up the risk of heart disease. The American College of Sport Medicine sets fitness categories for body fat based on age and gender, but there isn’t not an agreed upon target for what’s considered ideal. Using the ACSM standards, most Americans could be classified in the “poor fitness” category, as the average body fat among adults in the U.S. is 33%. The U.S. military considers the optimal body fat for military fitness to be between 10% and 20% for young men and up to 25% for middle-aged men. Women typically have more body fat, with an ideal range from 15% to 30% for young women and up to 38% for middle-aged women.

Millions of women are 'under-muscled.' These foods help build strength

Joseph says the rule of thumb for fitness is that “it’s you versus you.” Rather than fixate on an external benchmark, “it’s most important to look at trends over time,” in your body composition, he says.

It’s possible to lose fat and gain muscle, without losing any weight. This is what happened to Karen White , who is 59, and a certified health coach in Alexandria, Virginia. She’s gained about three pounds of muscle over the last three years, and has shed body fat,too. Her body fat has dropped from 26% down to 22%. “Literally, I’m the exact same weight,” after three years of tracking, but the positive changes in her body composition are profound.

Karen White is a certified health coach in Alexandria, Virginia. She credits body composition tests with helping her reduce body fat and build muscle. Jackie Cooke hide caption

She lifts weights three times a week, for about 30 minutes and has progressively built up to lifting heavier weights. She still does cardio work-outs and stays active with daily walks with her dog, though she has shifted her focus to resistance training.

“The misperception is that older people can't gain muscle, and that's absolutely not true,” she says. She points to a new client she’s working with in her 60s. Already, after a few months, her client has lost body fat and increased her strength.

White agrees that it’s important to set realistic expectations, and recognize the changes in body composition may take time. She has gained about a pound of muscle per year, on average, and feels a lot stronger.

Given that muscle peaks in our 30s, it’s important to do strength-training to maintain muscle mass, especially as we age. “The risk of frailty really increases exponentially with age,” Dr. Joseph says and muscle-loss, also called sarcopenia, affects an estimated 45% of older adults, especially women. Weight training can help fend off this loss.

Find Allison Aubrey on Instagram at  @allison.aubrey  and on X  @AubreyNPR .

This story was edited by Jane Greenhalgh

• body mass index
• Muscle strength

• Introduction
• Conclusions
• Article Information

The final cohort included participants who were older than 55 years with completed information to calculate body mass index (BMI) and aspirin frequency information. BQ indicates baseline questionnaire; SQ, supplemental questionnaire.

eTable 1. Univariable Proportional Hazards Models of Colorectal Cancer (CRC) and Non-CRC Gastrointestinal Cancer (GI) Incidence

eTable 2. Cancer Characteristics by Body Mass Index at Randomization

eTable 3. Multivariable Analysis of Non-CRC GI Cancer Risk (Liver, Pancreatic, Esophageal, Gastric) by Categorical Body Mass Index (BMI) at Early, Mid- and Later Adulthood

eTable 4. Colorectal and Non-Colorectal Gastrointestinal Cancer Risk in Early, Mid- and Later Adulthood, Using Body Mass Index (BMI) as a Continuous Variable

eTable 5. Multivariable Analysis of CRC and Non-CRC GI Cancer Risk by Early, Mid- and Later Adulthood Body Mass Index (BMI) Among Frequent Aspirin Users

Data Sharing Statement

• Obesity and Gastrointestinal Cancer: A Life Course Perspective JAMA Network Open Invited Commentary May 10, 2023 Mengyao Shi, MBBS, MPH; Yin Cao, ScD, MPH

See More About

Sign up for emails based on your interests, select your interests.

Customize your JAMA Network experience by selecting one or more topics from the list below.

• Academic Medicine
• Acid Base, Electrolytes, Fluids
• Allergy and Clinical Immunology
• American Indian or Alaska Natives
• Anesthesiology
• Anticoagulation
• Art and Images in Psychiatry
• Artificial Intelligence
• Assisted Reproduction
• Bleeding and Transfusion
• Caring for the Critically Ill Patient
• Challenges in Clinical Electrocardiography
• Climate and Health
• Climate Change
• Clinical Challenge
• Clinical Decision Support
• Clinical Implications of Basic Neuroscience
• Clinical Pharmacy and Pharmacology
• Complementary and Alternative Medicine
• Consensus Statements
• Coronavirus (COVID-19)
• Critical Care Medicine
• Cultural Competency
• Dental Medicine
• Dermatology
• Diabetes and Endocrinology
• Diagnostic Test Interpretation
• Drug Development
• Electronic Health Records
• Emergency Medicine
• End of Life, Hospice, Palliative Care
• Environmental Health
• Equity, Diversity, and Inclusion
• Facial Plastic Surgery
• Gastroenterology and Hepatology
• Genetics and Genomics
• Genomics and Precision Health
• Global Health
• Guide to Statistics and Methods
• Hair Disorders
• Health Care Delivery Models
• Health Care Economics, Insurance, Payment
• Health Care Quality
• Health Care Reform
• Health Care Safety
• Health Care Workforce
• Health Disparities
• Health Inequities
• Health Policy
• Health Systems Science
• History of Medicine
• Hypertension
• Images in Neurology
• Implementation Science
• Infectious Diseases
• Innovations in Health Care Delivery
• JAMA Infographic
• Law and Medicine
• Leading Change
• Less is More
• LGBTQIA Medicine
• Lifestyle Behaviors
• Medical Coding
• Medical Devices and Equipment
• Medical Education
• Medical Education and Training
• Medical Journals and Publishing
• Mobile Health and Telemedicine
• Narrative Medicine
• Neuroscience and Psychiatry
• Notable Notes
• Nutrition, Obesity, Exercise
• Obstetrics and Gynecology
• Occupational Health
• Ophthalmology
• Orthopedics
• Otolaryngology
• Pain Medicine
• Palliative Care
• Pathology and Laboratory Medicine
• Patient Care
• Patient Information
• Performance Improvement
• Performance Measures
• Perioperative Care and Consultation
• Pharmacoeconomics
• Pharmacoepidemiology
• Pharmacogenetics
• Pharmacy and Clinical Pharmacology
• Physical Medicine and Rehabilitation
• Physical Therapy
• Physician Leadership
• Population Health
• Primary Care
• Professional Well-being
• Professionalism
• Psychiatry and Behavioral Health
• Public Health
• Pulmonary Medicine
• Regulatory Agencies
• Reproductive Health
• Research, Methods, Statistics
• Resuscitation
• Rheumatology
• Risk Management
• Scientific Discovery and the Future of Medicine
• Shared Decision Making and Communication
• Sleep Medicine
• Sports Medicine
• Stem Cell Transplantation
• Substance Use and Addiction Medicine
• Surgical Innovation
• Surgical Pearls
• Teachable Moment
• Technology and Finance
• The Art of JAMA
• The Arts and Medicine
• The Rational Clinical Examination
• Tobacco and e-Cigarettes
• Translational Medicine
• Trauma and Injury
• Treatment Adherence
• Ultrasonography
• Users' Guide to the Medical Literature
• Vaccination
• Venous Thromboembolism
• Veterans Health
• Women's Health
• Workflow and Process
• Wound Care, Infection, Healing

Get the latest research based on your areas of interest.

Others also liked.

• Download PDF
• X Facebook More LinkedIn

Loomans-Kropp HA , Umar A. Analysis of Body Mass Index in Early and Middle Adulthood and Estimated Risk of Gastrointestinal Cancer. JAMA Netw Open. 2023;6(5):e2310002. doi:10.1001/jamanetworkopen.2023.10002

Manage citations:

© 2024

• Permissions

Analysis of Body Mass Index in Early and Middle Adulthood and Estimated Risk of Gastrointestinal Cancer

• 1 Division of Cancer Prevention and Control, Department of Internal Medicine, College of Medicine, The Ohio State University, Columbus
• 2 Comprehensive Cancer Center, The Ohio State University, Columbus
• 3 Gastrointestinal and Other Cancers Research Group, Division of Cancer Prevention, National Cancer Institute, Rockville, Maryland
• Invited Commentary Obesity and Gastrointestinal Cancer: A Life Course Perspective Mengyao Shi, MBBS, MPH; Yin Cao, ScD, MPH JAMA Network Open

Question   Is body mass index (BMI) or changing BMI associated with risk of gastrointestinal cancer?

Findings   This cohort study, which used data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, found that overweight and obese BMI in early and middle adulthood was associated with an increased risk of gastrointestinal cancer. Maintaining or increasing overweight or obese BMI over time was also associated with an increased risk of gastrointestinal cancer.

Meaning   These findings suggest that overweight and obese BMI over time may increase one’s risk of gastrointestinal cancer.

Importance   In a population with significantly increasing rates of individuals with overweight or obesity, understanding the association of obesity with long-term disease risk, such as cancer, is necessary to improve public health.

Objective   To investigate the association between body mass index (BMI) and gastrointestinal (GI) cancer risk (colorectal cancer [CRC] and noncolorectal GI cancer) in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial.

Design, Setting, and Participants   This retrospective cohort study was a secondary analysis of data from the PLCO Cancer Screening Trial. Participants aged 55 to 74 years were enrolled and randomized to the intervention (screening group) or control group at 10 screening centers between November 8, 1993, and July 2, 2001. The initial analysis of PLCO Cancer Screening Trial data occurred after 13 years of follow-up or December 31, 2009, whichever came first. Participants were reconsented in 2011 and either continued follow-up or refused additional follow-up. For those who reconsented, follow-up for incident cancers continued until December 31, 2014, or death, whichever occurred first. Data analysis for this secondary analysis was performed from April 2022 through November 2022.

Exposures   Body mass index and aspirin use, defined as the frequency of use of aspirin or aspirin-containing substances in the last 12 months.

Main Outcomes and Measures   The primary outcomes were the diagnoses of CRC and noncolorectal GI cancer. The association between BMI and cancer (CRC and noncolorectal GI cancer) was assessed using Cox proportional hazards regression modeling. The association between cancer risk and change in BMI was further analyzed at different ages, and an exploratory analysis was performed to evaluate GI cancer risk among aspirin users.

Results   This analysis included 135 161 participants (median [range] age, 62 [55-78] years; 67 643 [50.0%] female). Overweight BMI in early adulthood (hazard ratio [HR], 1.23; 95% CI, 1.10-1.37) and overweight BMI in middle adulthood (HR, 1.23; 95% CI, 1.13-1.34) and later adulthood (HR, 1.21; 95% CI, 1.10-1.32) as well as obese BMI in middle adulthood (HR, 1.55; 95% CI, 1.38-1.75) and later adulthood (HR, 1.39; 95% CI, 1.25-1.54) were associated with increased risk of CRC. Similar results were observed for the association with overall GI and non-CRC GI risk and BMI in middle and later adulthood. Maintaining overweight or obese BMI or increasing BMI to overweight or obese in later adulthood was also associated with increased CRC risk. Aspirin use 3 or more times per week did not significantly modify this association.

Conclusions and Relevance   In this secondary analysis of the PLCO Cancer Screening Trial, overweight and obese BMI in early and middle adulthood was associated with an elevated risk of CRC and noncolorectal GI cancers. The results of the current study prompt further exploration into the mechanistic role of obese BMI in carcinogenesis.

Colorectal cancer (CRC) is the third most incident cancer among men and women in the US. 1 Although improvements in CRC detection and screening have shifted CRC diagnosis to more localized and regional disease, a steadily decreasing but still staggering number of incident CRC cases are diagnosed annually. 1 This may be due to a concurrent increase in risk factors for gastrointestinal (GI) cancer development. Of particular interest, obesity rates are increasing globally. 2 Obesity is associated with numerous negative outcomes, including the development of type 2 diabetes and other metabolic disorders; cardiovascular diseases, such as hypertension and stroke; and cancer. 3 - 7 The World Cancer Research Fund and the International Agency for Cancer Research have estimated that approximately 20% of cancers may be attributed to excess weight gain. 8 - 10 Gastrointestinal cancers have been strongly associated with obesity, likely because of persistent, chronic inflammation attributable to obesity. 11 , 12 Chronic inflammation has been shown to be associated with increased risk of several GI cancers, such as pancreatic (pancreatitis), esophageal (esophagitis and Barrett esophagus), and colorectal (ulcerative colitis and Crohn disease). Epidemiological studies have consistently demonstrated increased GI cancer risk among individuals with overweight and obesity. 13 Furthermore, an analysis 14 of the Cancer Prevention Study II found that the risk of GI cancer–specific mortality increased 1.86 to 4.52 among men with obesity and 1.46 to 2.76 times among women with obesity compared with individuals with normal body mass index (BMI [calculated as weight in kilograms divided by height in meters squared]) (18.5-24.9–times increase).

The role of inflammation in cancer dates to observations by Virchow 15 and was expanded on by Dvorak. 16 Inflammation can be ascribed to several means, including chronic infection or conditions that result in enhanced proinflammatory signaling, such as obesity. However, many questions remain regarding the impact of heightened baseline inflammation attributed to obesity on cancer risk, such as the effect of obesity or weight gain in early life on later cancer risk or how changing BMI over time alters cancer risk. Recently, the DACHS (Darmkrebs: Chancen der Verhütung durch Screening [Colorectal Cancer: Chances for Prevention Through Screening]) study, a case-control study that evaluated risk factors and screening practices for CRC, found that obesity in early adulthood was associated with increased CRC risk, suggesting that early life events impact later health outcomes. 17 Previous assessments 18 , 19 of BMI in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial found an increased risk of mortality with a 5% or greater increase in BMI, and a separate analysis 20 found that adenoma and CRC risk was associated with increasing BMI trajectories. A meta-analysis 21 of prospective studies investigating BMI found a pooled relative risk of 1.33 (95% CI, 1.25-1.42) comparing obese with normal BMI and 1.46 (95% CI, 1.33-1.60) comparing highest with lowest categories of waist circumference. Although the literature has established the precedent that BMI influences CRC risk, much still needs to be explored. Therefore, in the current study, we evaluated the association between GI cancer, CRC, and noncolorectal GI cancer risk and BMI at early, middle, and later adulthood, as well as the association between changing BMI and cancer risk.

This cohort study was a secondary analysis of the PLCO Cancer Screening Trial, a large, multicenter randomized clinical trial that evaluated the efficacy of prostate, lung, colorectal, and ovarian cancer screening examinations in reducing mortality. The PLCO Cancer Screening Trial was approved by the institutional review boards of all study sites. Participants provided written informed consent for the original and ancillary studies. Additional approval for the current study was not required because data use in ancillary studies was included in the original consent and all data were deidentified. This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology ( STROBE ) reporting guideline. 22

The PLCO Cancer Screening Trial study design has been described elsewhere. 23 - 25 Briefly, participants aged 55 to 74 years were enrolled and randomized to the intervention (screening group) or control group at 10 screening centers (University of Alabama at Birmingham, Georgetown University, University of Pittsburgh, Washington University in St Louis, University of Utah, University of Colorado, University of Minnesota, Pacific Health Research and Education Institute [Hawaii], the Henry Ford Health System [Detroit, Michigan], and Marshfield Clinic Research Foundation [Marshfield, Wisconsin]) between November 8, 1993, and July 2, 2001. Exclusion criteria pertinent to the current analysis were age younger than 55 or older than 74 years at the time of randomization; a history of prostate, lung, colorectal, or ovarian cancer; prior surgical removal of the colon; treatment for cancer other than basal or squamous cell carcinoma of the skin; participation in another cancer screening or cancer primary prevention trial; beginning in April 1995, receipt of a colonoscopy, sigmoidoscopy, or barium enema in the 3 years before enrollment; and unwillingness or inability to sign a consent form. Additional PLCO Cancer Screening Trial exclusion criteria can be found elsewhere. 25 Participants randomized to the intervention group received screening for prostate, lung, colorectal, and ovarian cancers in the designated study years, whereas participants in the control group received standard care.

To complete the current analysis, we implemented additional exclusion criteria. Participants were excluded from the final analysis for the following reasons: (1) no valid baseline questionnaire (BQ), (2) BMI or aspirin use information was incomplete, (3) any personal history of cancer, and (4) discordant responses between the BQ and supplemental questionnaire (SQ), if the SQ was completed and valid and responses were used for analysis ( Figure ). A BQ was completed on or soon after enrollment. The SQ was distributed to study participants between 2006 and 2008, although completion of the SQ was not required. Both questionnaires are publicly available. 26 , 27 Age-specific BMI was calculated using self-reported height and weight at the designated ages from the BQ. In the current analysis, BMI at the age of 20 years is considered early adulthood, BMI at 50 years is considered middle adulthood, and BMI at the time of the study is considered later adulthood, as this refers to individuals 55 years or older. Specifically, the questions used to calculate BMI were as follows: “What is or was your weight at these ages? (Enter the weight in pounds.),” with response categories for weight at 50 years of age, 20 years of age, and current weight (at BQ); and “How tall are you? (Record your height in feet and inches.).” Body mass index was calculated and categorized according to the World Health Organization standard categorization: underweight (BMI <18.5), normal (BMI of 18.5-24.9), overweight (BMI of 25.0-29.9), and obesity (BMI ≥30). 28 The exploratory analysis used the following questions regarding aspirin use from the BQ and SQ, respectively: “During the last 12 months, have you regularly used aspirin or aspirin-containing products, such as Bayer, Bufferin, or Anacin (Please do not include aspirin-free products such as Tylenol or Panadol)?” and “During the last 12 months, about how often did you usually take aspirin (examples of aspirin include Bayer, Bufferin, Anacin, and baby aspirin)?” Participants who reported aspirin use 3 or more times per week were used for the final analysis, a threshold previously established. 29

The initial analysis of PLCO Cancer Screening Trial data occurred after 13 years of follow-up or December 31, 2009, whichever came first. 23 Participants reconsented in 2011 and either continued follow-up or refused additional follow-up. For those who reconsented, follow-up for incident cancers continued until December 31, 2014, or death, whichever occurred first. If the incident cancer occurred after reconsent, incident cancers were determined by state cancer registry linkage. The mean (SD) follow-up of the included cohort was 13.9 (6.0) years, and the median (range) follow-up was 14.9 (0-24.2) years. Data analysis for this secondary analysis was performed from April 2022 through November 2022.

The goal of this analysis was to evaluate the association between BMI and GI cancer risk, separating CRC or noncolorectal GI cancer incidence in the PLCO Cancer Screening Trial. We defined an incident cancer as the first cancer diagnosed during follow-up. Diagnostic information on incident cancers was collected and recorded using an approved medical record abstraction form. International Classification of Diseases for Oncology, Second Edition ( ICD-O-2 ) codes and diagnosis dates were collected for nonstudy cancer, whereas additional information, such as stage and grade, was collected for study cancers. Incident GI cancers were CRC (codes 153 and 154), esophageal (code 150), gastric (code 151), liver (code 155), and pancreatic (code 157) cancers, identified by International Classification of Diseases, Ninth Revision ( ICD-9 ) codes ( Figure ). Follow-up time began at the time of randomization and continued until the date of cancer diagnosis, participant death, or the end of study follow-up. Individuals diagnosed with cancer not included in a given analysis were censored at the time of diagnosis to account for competing risks.

Time-dependent Cox proportional hazards regression models with competing risks were used to calculate hazard ratios (HRs) and 95% CIs, assessing the associations between BMI and GI cancer (CRC and noncolorectal GI). Univariable regression modeling for risk of CRC and noncolorectal GI cancer was performed to determine variables for inclusion in the multivariable model (eTable 1 in Supplement 1 ). Variables with a univariable P  < .05 were included in the final model. Covariates included in the final regression model were age at randomization, study randomization group (intervention or control), study center (University of Colorado, Georgetown University, Pacific Research and Education Institute [Hawaii], Henry Ford Health System, University of Minnesota, Washington University in St Louis, University of Pittsburgh, University of Utah, Marshfield Clinic Research Foundation [Wisconsin], or University of Alabama at Birmingham), sex (male or female), self-reported race and ethnicity (Asian, Hispanic, or Pacific Islander [grouped because the number of participants for each group was small and not appropriately powered to generate an estimate], non-Hispanic Black, and non-Hispanic White), smoking status (never, current, or former), ibuprofen use (<3 or ≥3 times per week), aspirin use (0-<1 time per month, 1-3 times per month, 1-2 times per week, or ≥3 times per week), and history of myocardial infarction, stroke, hypertension, or diabetes. In univariable analyses, race and ethnicity were statistically significant and therefore included in the model. Smoking status, ibuprofen use, aspirin use, and history of myocardial infarction, stroke, hypertension, and diabetes were incorporated into the models as time dependent. The Cochran-Armitage test was used to test for the underlying pattern between the BMI categories. All statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc). P values were 2-tailed, and statistical significance was set at P  < .05.

Of 154 887 participants enrolled in the PLCO Cancer Screening Trial, 135 161 (median [range] age, 62 [55-78] years; 8726 [6.5%] Asian, Hispanic, or Pacific Islander; 6920 [5.1%] non-Hispanic Black; 119 453 [88.4%] non-Hispanic White; 62 [0.1%] missing race; 67 643 [50.0%] female and 67 518 [50.0%] male) were included in the analysis. The mean (SD) follow-up time of the eligible cohort was 13.9 (6.0) years, and the median (range) was 14.9 (0-24.2) years. During follow-up, 34 946 (25.9%) were diagnosed with cancer, with 5088 (14.6%) GI cancers. A total of 2803 (55.1%) of the incident GI cancers were CRC, with 376 esophageal cancers (7.4%), 485 gastric cancers (9.5%), 348 liver cancers (6.8%), and 1076 pancreatic cancers (21.1%). The demographic characteristics of the analyzed cohort are included in Table 1 , and cancer characteristics are given in eTable 2 in Supplement 1 . We modeled BMI at (1) early adulthood (BMI at 20 years of age), (2) middle adulthood (BMI at 50 years of age), and (3) later adulthood (BMI at ≥55 years of age) as both categorical and continuous variables to evaluate the association between BMI and GI cancer risk. Increased risk of overall GI cancer was observed among individuals with overweight (early adulthood: HR, 1.17; 95% CI, 1.08-1.27; middle adulthood: HR, 1.18; 95% CI, 1.11-1.26; later adulthood: HR, 1.17; 95% CI, 1.09-1.25) and obesity (early adulthood: HR, 1.31; 95% CI, 1.08-1.59; middle adulthood: HR, 1.50; 95% CI, 1.37-1.64; later adulthood: HR, 1.38; 95% CI, 1.27-1.49) in early, middle, and later adulthood ( Table 2 ). We observed an increased risk of CRC for individuals with overweight BMI (HR, 1.23; 95% CI, 1.10-1.37) in early adulthood, overweight (HR, 1.23; 95% CI, 1.13-1.34) and obese (HR, 1.55; 95% CI, 1.38-1.75) BMI in middle adulthood, and overweight (HR, 1.21; 95% CI, 1.10-1.32) and obese (HR, 1.39; 95% CI, 1.25-1.54) BMI in later adulthood. Similarly, increased risk of noncolorectal GI cancer was associated with obese BMI (HR, 1.37; 95% CI, 1.04-1.80) in early adulthood, overweight (HR, 1.13; 95% CI, 1.03-1.24) and obese (HR, 1.44; 95% CI, 1.27-1.65) BMI in middle adulthood, and overweight (HR, 1.13; 95% CI, 1.03-1.24) and obese (HR, 1.36; 95% CI, 1.21-1.53) BMI in later adulthood. Individual non-GI cancer risk estimates are included in eTable 3 in Supplement 1 . When modeled continuously, we observed 2% to 4% increased risk of both CRC and noncolorectal GI cancer with each 1-unit increase in BMI across all time points (eTable 4 in Supplement 1 ).

We next wanted to investigate whether changing BMI over time differentially influenced CRC and noncolorectal GI cancer risk, particularly if the participants changed BMI categories, for example, moving from normal BMI in early adulthood to overweight in later adulthood ( Table 3 ). We found that individuals who had overweight or obese BMI in early and later adulthood (HR, 1.45; 95% CI, 1.28-1.64; P  < .001) and those who moved from underweight or normal BMI in early adulthood to overweight or obese BMI in later adulthood (HR, 1.23; 95% CI, 1.13-1.34; P  < .001) had increased risk of CRC. A similar pattern was observed for noncolorectal GI cancer risk (no change in overweight or obese BMI: HR, 1.29; 95% CI, 1.13-1.47; P  < .001; underweight or normal to overweight or obese BMI: HR, 1.17; 95% CI, 1.06-1.29; P  = .002). When we modeled change in BMI from middle to later adulthood, we found that consistent overweight or obese BMI (HR, 1.37; 95% CI, 1.25-1.51; P  < .001), changing from overweight or obese to underweight or normal BMI (HR, 1.47; 95% CI, 1.21-1.78; P  < .001), and changing from underweight or normal to overweight or obese BMI (HR, 1.20; 95% CI, 1.06-1.34; P  = .003) were associated with increased risk of CRC. Statistical significance was only observed for static overweight or obesity between middle and later adulthood and noncolorectal GI cancer risk (HR, 1.25; 95% CI, 1.12-1.38; P  < .001). These data suggest that alterations in BMI over time may influence one’s risk of GI cancer.

Finally, because of the potential influence of overweight or obese BMI on cancer preventive agent efficacy, we wanted to evaluate the association between BMI and CRC and noncolorectal GI cancer risk among frequent aspirin users. Among frequent aspirin users, overweight or obese BMI in early (HR, 1.44; 95% CI, 1.23-1.68; P  < .001), middle (HR, 1.45; 95% CI, 1.26-1.66; P  < .001), and later (HR, 1.43; 95% CI, 1.24-1.65; P  < .001) adulthood was associated with increased risk of CRC (eTable 5 in Supplement 1 ). Similar associations were observed with noncolorectal GI cancer risk.

In this cohort study, we found that overweight and obese BMI at different ages and change in BMI over time may be associated with increased risk of GI cancers. We further found that aspirin use 3 or more times per week did not modify this association. Aspirin use for cancer prevention has been well supported by decades of epidemiological evidence. Previous secondary analyses 29 , 30 demonstrated the efficacy of aspirin in reducing the risk of CRC and bladder cancer mortality. However, the impact of BMI on this association has not been adequately delineated. Furthermore, the updated US Preventive Services Task Force recommendations for aspirin use to prevent cardiovascular disease discusses the withdrawal of aspirin use for CRC prevention, of which it had previously been given a B rating for individuals aged 50 to 69 years with a 10% or greater risk of cardiovascular disease, citing insufficient or conflicting evidence. 31 - 33

Obesity results from the buildup and storage of white adipose tissue, or fat. Adipose cells can induce the inflammatory response and promote immune cell dysfunction through the secretion of adipokines and proinflammatory cytokines, leading to further downstream mechanistic dyregulation. 34 Individuals with obesity are at higher risk of several conditions, including cancer. Interestingly, not all cancers are significantly associated with obesity; rather, it is more limited to those where cancer cells grow near adipose cells, potentially due to the impact of adipose cells on tumorigenesis. 35 Research has indicated significant crosstalk between cancer cells and adipocytes. For example, in vitro CRC cell line coculture with adipocytes has demonstrated increased cancer cell proliferation, migration, and nutrient transfer (eg, ketones and fatty acids) from adipocytes to the cancer cells. 36 Transcriptomic analysis of the ColoCare Study, a prospective cohort of newly diagnosed CRCs, found enrichment of pathways, such as fibrosis and glycolytic metabolism, associated with adipose–tumor tissue crosstalk. 37 Similar findings have been observed for noncolorectal GI cancers. 38 - 40 Although likely not the initiator, excess adipocytes promote tumorigenesis through supplying cancer cells with much-needed nutrients and stimulation of oncogenic pathways. Therefore, cancer prevention mechanisms that target the harmful physiologic effects of obesity may work to counteract tumorigenesis.

As found in the current study, obesity may alter the cancer preventive effect of aspirin. Our results indicate that individuals with overweight and obese BMIs had an increased risk of CRC and noncolorectal GI cancer with aspirin use 3 or more times per week, suggesting that aspirin may not be efficacious for prevention in overweight or obese states. The ability of aspirin to protect against GI cancers may be blunted in people with obesity because of inadequate dosing. 41 , 42 A suggestion may be that individuals with obesity need to increase aspirin frequency or dosage; however, increased aspirin use comes with its own risks, such as gastrointestinal bleeding. 43 In our analysis, we did not account for participant dosing, a noted limitation to the study. Additional studies evaluating the impact of aspirin dose on cancer prevention, accounting for participant BMI or weight gain, are needed to better delineate aspirin’s role. The Cancer Prevention Project 3 (CaPP3) is currently under way to discern the effect of differential aspirin dosing (100, 300, or 600 mg) in a cohort of individuals with Lynch syndrome. The CaPP3 study is ongoing; however, the eventual results of this study may be translatable to the general, average-risk population.

Although the findings from the current study are significant, important limitations should be noted. Despite the strengths of baseline and supplemental information, including detailed BMI information at different ages and extended follow-up with linked participant outcomes, the current study is a secondary analysis of a completed cancer screening trial; therefore, the collection of exposure and outcome information was not included as part of the original study. Additionally, all BQ and SQ information was collected by self-report, which includes height and weight data used to calculate BMI. Therefore, aspirin dosing information was not collected as part of the BQ and was not accounted for in this analysis. In addition, we were unable to correlate changes in BMI with aspirin use. Aspirin use, as stated in the BQ and SQ, was reported during the last year, whereas BMI could have changed at any point before the questionnaires. However, in the current analysis, we were only evaluating the association between these 2 factors, not causation. Finally, it is possible that not all confounders were accounted for in our multivariable logistic regression models.

This cohort study found increased GI cancer risk among individuals with overweight and obese BMI reported at early, middle, and later adulthood. We also found that increasing BMI over time was associated with increased risk of CRC and noncolorectal GI cancers. This association was not modified by aspirin use 3 or more times per week. The results of the current study prompt further exploration into the mechanistic role of obese BMI in carcinogenesis. Finally, future research must focus on identifying cancer prevention mechanisms for this high-risk group.

Accepted for Publication: March 10, 2023.

Published: May 10, 2023. doi:10.1001/jamanetworkopen.2023.10002

Open Access: This is an open access article distributed under the terms of the CC-BY License . © 2023 Loomans-Kropp HA et al. JAMA Network Open .

Corresponding Author: Holli A. Loomans-Kropp, PhD, MPH, Division of Cancer Prevention and Control, Department of Internal Medicine, The Ohio State University, 1590 N High St, Ste 571, Columbus, OH 43201 ( [email protected] ).

Author Contributions: Dr Loomans-Kropp had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Loomans-Kropp.

Acquisition, analysis, or interpretation of data: Both authors.

Drafting of the manuscript: Loomans-Kropp.

Critical revision of the manuscript for important intellectual content: Both authors.

Statistical analysis: Loomans-Kropp.

Supervision: Umar.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported in part by funds from The Ohio State University (Dr Loomans-Kropp) and intramural funding from the National Institutes of Health (Drs Loomans-Kropp and Umar).

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2 .

Additional Information: Cancer incidence data have been provided by the Colorado Central Cancer Registry, District of Columbia Cancer Registry, Georgia Cancer Registry, Hawaii Cancer Registry, Cancer Data Registry of Idaho, Minnesota Cancer Surveillance System, Missouri Cancer Registry, Nevada Central Cancer Registry, Pennsylvania Cancer Registry, Texas Cancer Registry, Virginia Cancer Registry, and Wisconsin Cancer Reporting System. All are supported in part by funds from the Centers for Disease Control and Prevention, National Program for Central Registries, local states, or the National Cancer Institute’s Surveillance, Epidemiology, and End Results program. The results reported here and the conclusions derived are the sole responsibility of the authors.

• Register for email alerts with links to free full-text articles
• Access PDFs of free articles
• Manage your interests
• Save searches and receive search alerts

• Health Tech
• Health Insurance
• Medical Devices
• Gene Therapy
• Neuroscience
• H5N1 Bird Flu
• Health Disparities
• Infectious Disease
• Mental Health
• Cardiovascular Disease
• Chronic Disease
• Alzheimer's
• Coercive Care
• The Obesity Revolution
• The War on Recovery
• Adam Feuerstein
• Matthew Herper
• Jennifer Adaeze Okwerekwu
• Ed Silverman
• CRISPR Tracker
• Breakthrough Device Tracker
• Generative AI Tracker
• Obesity Drug Tracker
• 2024 STAT Summit
• Wunderkinds Nomination
• STAT Madness
• STAT Brand Studio

Don't miss out

Subscribe to STAT+ today, for the best life sciences journalism in the industry

Study finds BMI, despite flaws, is useful estimate of body fat in children

By Elizabeth Cooney June 3, 2024

T he body mass index has long been slammed as a blunt instrument for evaluating health, even more so with new obesity drugs changing the conversation about weight and well-being. Now a study reasserts BMI’s value as a screening tool in children to detect high levels of body fat, a measure tied to greater risk of cardiovascular disease, early atherosclerosis, and a high BMI in adulthood.

BMI is an equation that divides a person’s weight in kilograms by their height in meters squared. For children, growth curves from the Centers for Disease Control and Prevention are used to track a child’s trajectory, rather than assign them to one of the four categories familiar to adults (underweight, healthy weight, overweight, obesity). For people of all ages, BMI is a proxy for body fat, whose best measurement involves dual-energy x-ray absorptiometry (DXA), an expensive tool impractical in primary care. 

advertisement

The new study recognizes the limitations of BMI because its two measures are simply height and weight, numbers that may be better suited to studying populations than to directing care for individual patients. And it notes that at the same BMI, Black children have less adiposity than white children. The American Medical Association has urged doctors to de-emphasize the use of BMI in assessing health and obesity, decrying its use for “racist exclusion” and for causing “historical harm” because it was based only on white populations.  

Critics also say those two metrics can’t distinguish fat mass from lean mass, so the researchers set out to see how well BMI compares to DXA when looking at the two kinds of body mass. They analyzed data using both measures from 6,923 young people age 8 through 19 in the National Health and Nutrition Survey, conducted from 2011 through 2018. 

Young people with a high BMI — defined as equal to or higher than the 95th percentile on the growth curve — were 29 times more likely than those with lower BMI to have a high fat mass. 

Because high BMI was more strongly linked to high levels of fat mass, its value would be greatest in screening for high adiposity, the authors said. 

Ihuoma Eneli, visiting professor of pediatrics-nutrition at the University of Colorado School of Medicine, called the study carefully done and much needed, given the controversies surrounding BMI as the core measure to define obesity. 

“We need to take criticisms seriously and see it as an opportunity to test and retest our assumptions, as this paper has done,” she told STAT via email. “It’s particularly important in pediatrics as the growth and development is one of the core areas we use to define a thriving child and features prominently in every aspect of pediatric practice/discipline. BMI use is not all about obesity alone.”

As a clinician, Eneli has had patients and parents question a high BMI number in a child who was an athlete. Sometimes she shared their doubt, but later testing would confirm excess body fat. 

“High BMI is a very good indicator of high fat mass,” senior study author David Freedman told STAT. Now retired from the CDC’s Division of Nutrition, Physical Activity, and Obesity, his work over 35 years concentrated on BMI and cardiovascular disease risk factors in children. 

“Throughout my career, there’s always been criticisms of BMI,” Freedman said. Although the study concluded BMI is not perfect, it has “utility” in research and medical care. “I wrote this study just to try to examine how good is BMI as an indicator of high adiposity,” he said.

Sign up for Morning Rounds

Understand how science, health policy, and medicine shape the world every day

BMI was better at its higher end in detecting high body fatness. When looking at children and adolescents at or above the 95th percentile for BMI, it’s important to remember that the CDC growth charts were drafted roughly 50 years ago. About 20% of children are now above that 95th percentile.  

The paper, posted Monday in Pediatrics , follows last year’s guidelines issued by the American Academy of Pediatrics (which publishes the scientific journal) recommending weight-loss drugs for children as young as 12 whose weight and age place them at the high end of growth charts. That policy, which acknowledged debate about BMI, drew pushback from experts concerned about focusing on weight, not health. For adults, critics have pointed to a 2016 Nature study that documented how relying on BMI alone could incorrectly classify both good and bad cardiometabolic health.

Late last year, the U.S. Preventive Services Task Force said it needed more evidence to fully understand the long-term health outcomes for weight-loss medications in children before making a recommendation. Instead, it advised clinicians to provide or refer children 6 or older who have a high BMI to intensive behavioral interventions to help them achieve a healthy weight and improve their quality of life. 

“Although BMI is not a perfect tool by any means, it is a simple tool which can be used to screen for adiposity and associated risks of adverse health outcomes,” said Sharon Weston, a senior clinical nutrition specialist at the Optimal Wellness for Life Clinic at Boston Children’s Hospital. She was not involved in the Pediatrics study. “It is important, however, to use BMI in conjunction with other parameters to measure overall health and wellness,” such as diet, sleep, and physical activity.

Related: Here’s why obesity grew so quickly worldwide, and where that’s starting to change

The new study did not make race and ethnicity comparisons, but cites other research based on the same data set that saw differences mostly at lower percentiles on the CDC BMI growth charts. Reasons for these disparities aren’t known, but the authors suggest environmental, social, behavioral, and nutritional factors may be at play. 

Eneli, who was a co-author of the American Academy of Pediatrics guidelines on obesity drugs in children, called the history of how BMI came about disconcerting. “One wonders what this study will look like analyzed using race/ethnicity,” she said .

Freedman said he has launched a study of those differences and expects to publish results next year.

Alternatives to BMI for screening include measuring waist and neck circumference, Weston said, but for their use to become standard in clinical practice, developing optimal cutoff points in children would need to be developed.

Today’s findings reinforce BMI’s usefulness for spotting children whose fat mass is rising, Jaime Moore and Stephen Daniels of Children’s Hospital Colorado wrote in a companion commentary , but it’s only a start.

“Improved standardization for the identification of pediatric obesity and severe obesity using BMI, when paired with equitable delivery of treatment, could help to counteract weight bias and reduce disparities in obesity-related health outcomes,” they said. “We remind practitioners that the standardized use of BMI to identify patients with obesity is a first step.”  

STAT’s coverage of chronic health issues is supported by a grant from  Bloomberg Philanthropies . Our financial supporters  are not involved in any decisions about our journalism.

About the Author Reprints

Elizabeth cooney.

Cardiovascular Disease Reporter

Elizabeth Cooney is a cardiovascular disease reporter at STAT, covering heart, stroke, and metabolic conditions.

cardiovascular disease

children's health

chronic disease

STAT encourages you to share your voice. We welcome your commentary, criticism, and expertise on our subscriber-only platform, STAT+ Connect

To submit a correction request, please visit our Contact Us page .

Recommended

Recommended Stories

NIH will bring clinical research into primary care offices with $30 million pilot

Smoke exposure from California’s wildfires linked to 52,000 early deaths, study says

STAT Plus: Elevance PBM’s president out as customers complain of prescription chaos

STAT Plus: Which hospital systems will join Kaiser’s Risant next?

STAT Plus: FDA advisory panel votes overwhelmingly against MDMA therapy for PTSD

This website is intended for healthcare professionals

• { $refs.search.focus(); })" aria-controls="searchpanel" :aria-expanded="open" class="hidden lg:inline-flex justify-end text-gray-800 hover:text-primary py-2 px-4 lg:px-0 items-center text-base font-medium"> Search

Search menu

Burr V, 3rd edn. London: Routledge; 2015

Catalano P, Presley L, Mnimum J, Haugal-de-Mouzon S Fetuses of obese mothers develop insulin resistance in utero. Diabetes Care. 2009; 32:(96)1076-1080 https://doi.org/10.2337/dc08-2077

Chan SW, Tulloch E, Cooper ES, Smith A, Wojcik W, Norman JE Montgomery and informed consent: where are we now?. BMJ. 2017; https://doi.org/10.1136/bmj.j2224

Chief Nursing Officers of England, Northern Ireland, Scotland and Wales. Midwifery 2020: delivering expectations. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/216029/dh_119470.pdf (accessed 26 November 2020)

Denison FC, Norrie G, Graham B, Lynch N, Reynolds RM Increased BMI is associated with an increased risk of minor complications during pregnancy with consequent cost implications. BCOG. 2009; 116:(11)1467-1472 https://doi.org/10.1111/j.1471-0528.2009.02222.x

Denison FC: NHS Lothian; 2011

Denison FC, Chiswick C Improving pregnancy outcome in obese women. Proceedings of the Nutrition Society. 2011; 11:(70)457-464

Drake A J, Reynolds RM Impact of maternal obesity on offspring obesity and cardiometabolic disease risk. Reproduction. 2010; 9:(140)387-398 https://doi.org/10.1530/REP-10-0077

Dunkley-Bent J, 15th edn. In: Henderson C, Macdonald S (eds). Edinburgh: Bailliere Tindall; 2004

Foster CE, Hirst J Midwives' attitudes towards giving weight related advise to obese pregnant women. British Journal of Midwifery. 2012; 24:(4)254-262 https://doi.org/10.12968/bjom.2014.22.4.254

Frain J(ed). UK: Wiley Blackwell; 2018

Franklin C Expanding the vision of the social constructionist debates: creating relevance for practitioners. Families in Society: The Journal of Contemporary Social Services. 1995; 76:(7)395-402

Furness PJ, McSeveny K, Arden MA, Garland C, Dearden AM, Soltani H Maternal obesity support services: a qualitative study of the perspectives of women and midwives. BMC Pregnancy and Childbirth. 2011; 11 https://doi.org/10.1186/1471-2393-11-69

Gergen KJ, 3rd edn. London: Sage Publications; 2015

Greenhalgh T, 5th edn. London: Sage Publications;

Gibbs GLos Angeles: Sage Publications; 2007

Haby K, Glantz A, Hanas R, Premberg A Mighty mums – an antenatal health care intervention can reduce gestational weight gain in women with obesity. Midwifery. 2015; 31:(7)685-692 https://doi.org/10.1016/j.midw.2015.03.014

Harding J, 2nd edn. Los Angeles: Sage Publications; 2019

Herring SJ, Platek DN, Elliott P, Riley LE, Stuebe AM, Oken E Addressing obesity in pregnancy: what do obstetric providers recommend?. Journal of Women's Health. 2010; 19:(1)65-70 https://doi.org/10.1089/jwh.2008.1343

Heslehurst N, Moore H, Rankin J, Ells L J, Wilkinson J R, Summerbell CD How can maternity services be developed to effectively address maternal obesity? A qualitative study. Midwifery. 2011; 27:170-177 https://doi.org/10.1016/j.midw.2010.01.007

Heslehurst NS, McCormack S, Sedgewick G, Bell R, Rankin J Mdwives perspectives of their training and education requirements in maternal obesity: a qualitative study. Midwifery. 2013; 29:(7)736-744 https://doi.org/10.1016/j.midw.2012.07.007

Jewell K, Avery A, Barber J, Simpson S The healthy eating and lifestyle in pregnancy (help) feasibility study. British Journal of Midwifery. 2014; 22:(10)727-736 https://doi.org/10.12968/bjom.2014.22.10.727

Jomeen J Psychological context of childbirth, 15th edn. In: Macdonald S, Johnson G (eds). Edinburgh: Elsevier; 2017

Keely A, Gunning M, Denison FC Maternal obesity in Pregnancy: Women's understanding of risks. British Journal of Midwifery. 2011; 19:(6)364-369 https://doi.org/10.12968/bjom.2011.19.6.364

Knight-Agarwal CR, Kaur M, Williams LT, Davey R, Davis D The views and attitudes of health professionals providing antenatal care to women with a high BMI: a qualitative research study. Women and Birth. 2014; 27:(1)138-144 https://doi.org/10.1016/j.wombi.2013.11.002

McParlin C, Bell R, Robson SC, Muirhead CR, Araujo-Soares V. What helps or hinders midwives to implement physical activity guidelines for obese pregnant women? A questionnaire survey using the theoretical domains framework. Midwifery. 2017; 49:110-116 https://doi.org/10.1016/j.midw.2016.09.015

Miles MB, Huberman AM, 2nd edn. London: Sage Publications; 1994

MBRRACE-UK. Mothers and babies reducing risk through audits and confidential enquiries. 2014. https://www.npeu.ox.ac.uk/mbrrace-uk (accessed 5 October 2019)

MBRRACE-UK Mothers and babies reducing risk through audits and confidential enquiries. 2018. https://www.npeu.ox.ac.uk/downloads/files/mbrrace-uk/reports/MBRRACE-UK%20Maternal%20Report%202018%20-%20Web%20Version.pdf (accessed 26 November 2020)

McGiveron A, Foster S, Pearce J, Taylor MA, McMullen S, McParlin C, Bell R, Robson SC, Muirhead CR, Araujo-Soares V What helps or hinders midwives to implement physical activity guidelines for obese pregnant women? A questionnaire survey using the theoretical domains framework. Midwifery. 2017; 49:110-116

NHS. Better Births: improving outcomes of maternity services in England – A five year forward view for maternity care. 2017. https://www.england.nhs.uk/wp-content/uploads/2016/02/national-maternity-review-report.pdf (accessed 9 December 2020)

Nursing and Midwifery Council. The code. 2018. https://www.nmc.org.uk/globalassets/sitedocuments/nmc-publications/nmc-code.pdf (accessed 26 November 2020)

Pan SY, Dixon L, Paterson H, Campbell N Increased BMI during pregnancy: how to midwife lead maternity carers respond?. Womens Health. 2015; 11:(4)461-469 https://doi.org/10.2217/2FWHE.15.8

NHS Quality Improvement Scotland. Pathways in maternity care. 2009. http://www.healthcareimprovementscotland.org/our_work/reproductive%2C_maternal_child/programme_resources/keeping_childbirth_natural.aspx#:~:text=Pathway%20for%20Normal%20Maternity%20Care&text=The%20KCND%20pathway%20was%20designed,accessing%20maternity%20care%20across%20Scotland (accessed 13 April 2021)

Ronnberg AK, Ostlund J, Fadl H, Gottvall T, Nilsson K Intervention during pregnancy to reduce excessive gestational weight gain—a randomised controlled trial. BJOG. 2015; 122:(4)537-544 https://doi.org/10.1111/1471-0528.13131

A growing problem. Does weight matter?. 2010. https://www.rcm.org.uk/news-views-andanalysis/news/maternalobesity-figures-reveal-growing-problem (accessed 31 July 2020)

Scottish Government. Scottish Multiple Deprivation Index. 2020. https://www.gov.scot/collections/scottish-index-of-multiple-deprivation-2020/ (accessed 31 July 2020)

Silverman J, Kurrz S, Draper J, 3rd edn. London: Radcliffe Publishing; 2013

Smith DM, Cooke A, Lavender T Maternal obesity is the new challenge; a qualitative study of health professionals' views towards suitable care for pregnant women with a body mass index (BMI) >=30 kg/m2. BMC Pregnancy and Childbirth. 2012; 12 https://doi.org/10.1186/1471-2393-12-157

Stirrat LI, Reynolds RM Effects of maternal obesity on early and long-term outcomes for offspring. Research and Reports in Neonatology. 2014; 4:43-53 https://doi.org/10.2147/RRN.S46783

Increased food energy supply as a major driver of the obesity epidemic: a global analysis. 2015. https://www.who.int/bulletin/volumes/93/7/14-150565/en/ (accessed 26 November 2020)

Walsh D, Downe S Appraising the quality of qualitative research. Midwifery. 2006; 22:(2)108-119

Willcox JC, Campbell KJ, Van DP, Hoban E, Pidd D, Wilkinson S Excess gestational weight gaIn: an exploration of midwives' views and practice. BMC Pregnancy and Childbirth. 2012; 12

World Health Organization. Obesity and overweight. 2015. https://www.who.int/en/news-room/fact-sheets/detail/obesity-and-overweight (accessed 26 November 2020)

World Health Organization. Obesity and overweight. 2018. https://www.who.int/en/news-room/fact-sheets/detail/obesity-and-overweight (accessed 26 November 2020)

Obesity matters: the skills that strengthen midwifery practice when caring for obese pregnant women

Yvonne Greig

Midwifery Lecturer, Edinburgh Napier University

View articles · Email Yvonne

Anne F Williams

Senior Nurse Lecturer, Queen Margaret University, Edinburgh

View articles

Margaret Coulter-Smith

Obese pregnant women (BMI>30 kg/m²) are at an increased risk of developing complications during pregnancy, labour and birth. Furthermore, their offspring are at risk of short- and long-term health complications. Midwives are ideally situated to inform women about risks and to support them in optimising their health. 

Obesity has been recognised as one of the leading health concerns in our society ( World Health Organization [WHO], 2018 ). Evidence suggests that good health is at risk for those individuals who live with BMI>30 kg/m² ( Herring et al, 2010 ; Russell et al, 2010 ; Furness et al, 2011 ; Heslehurst et al, 2013 ; Foster et al, 2014). However, obesity rates continue to rise, and obesity remains prevalent both nationally and globally ( WHO, 2015 ; Scottish Public Health Observatory, 2018), multi-factorial reasons have been cited for this such as diets high in obesogenic foods and sedentary lifestyles (Vandevijvere et al, 2018).

Living with a raised BMI>30 kg/m² while pregnant carries risks for both women and their babies that can lead to poor pregnancy outcomes for both and risk the ongoing health of the off spring ( Catalano et al, 2009 ; Denison et al, 2009 ; Denison and Chiswick 2011 ; Drake and Reynolds, 2014; Stirrat and Reynolds, 2014 ; MBRRACE-UK 2014 ; 2018 ; Keely, 2015). Evidence suggests that the continued prevalence of obesity in pregnant women will impact upon future generations, who will, in turn, experience the incumbent risks of obesity and sub-optimal health. It is likely, therefore, that the ongoing cycle of obesity will continue to impact on the costs of future healthcare provision.

Midwives, as the first professional point of contact for many women, have a key role to play in providing appropriate evidenced-based advice regarding lifestyle choices of which diet, physical activity and weight management are a few ( Dunkley-Bent, 2004 ). Furthermore, providing women with dietary and physical activity advice has been found to improve maternal health and pregnancy outcomes ( Jewell et al, 2014 ; McGiveron et al, 2014; Ronnberg et al, 2014; Haby et al, 2015 ). Raising the topic of and maintaining dialogue with women about obesity has been found to be challenging, with professionals being fearful of offending women ( Furness et al, 2011 ; Macleod et al, 2013; Wilkinson et al, 2013; Foster and Hirst, 2014). Research was undertaken to gain deeper understanding of community midwifery practice and how appropriate information is provided to obese women in South East Scotland.

Antenatal care in South East Scotland

Community midwives who took part in this study are based in various health centres and GP surgeries, and ensure that every woman in every geographical area has access to midwifery services. During the antenatal course, the community midwife is responsible for providing holistic care to women, ensuring that appropriate referrals are made to specialists when risks are identified or deviations from normal occur, ensuring appropriate care is provided ( NHS Quality Improvement Scotland, 2009 ; NHS, 2017 ).

Aim of research study

The aim of the research was to gain deeper understanding about how community midwives practice with respect to raising the topic of obesity and to explore how they maintained dialogue with women about this. Community midwives are well placed to discuss complex health issues with women and their families throughout pregnancy and in the postnatal period of which living with a raised BMI is one. Prior to commencing the research, a narrative literature review was undertaken, the findings of which informed the study design.

Literature review

The literature search aimed to identify literature that reported on professional practice with respect to caring for and advising obese pregnant women. Figure 1 outlines the search strategy that was utilised during this review. The literature was reviewed using questions adapted from Greenhalgh (1998) for quantitative papers, and Walsh and Downe (2006) for qualitative papers. A total of three dominant themes were found ( Table 4 ); midwives perceived that they lacked proficiency in advice giving, midwives were fearful of damaging the midwife/woman relationship, midwives had received insufficient education (about obesity as a topic).

Ethical approval

Ethics approval for this study was gained from the Higher Education Institution (HEI) and Research and Development department of the NHS Health Board (R&D 2017/0316) where the investigation was to be undertaken, NHS ethical approval for this study was not required because no service users were asked to participate.

Recruitment

Initially, face-to-face meetings were held with community midwifery teams, information sheets were distributed to individuals for them to consider and a follow-up call was made to them after a 24-hour period asking if they would like to participate. Extra information sheets were left in each office to allow those who had been off duty on the day of the visit, ensuring that the research was effectively advertised. Two midwives made contacted with the researcher as a result of this strategy and participated in the study. In total, 13 practicing community midwives consented to take part. A summary of participant characteristics can be seen in Table 1 .

Data collection

Data were collected from January to May 2018. Following written and ongoing consent, in-depth interviews were chosen as the data collection method and midwives were also provided with a practice diary asking them to reflect upon up to five episodes where they had provided care to women who had a BMI≥30 kg/m². Interviews lasted between 20–65 minutes. The interview schedule was informed by the initial narrative literature review focusing upon professional practice and advice giving in the context of caring for obese pregnant women ( Herring et al, 2010 ; Stotland et al, 2010; Smith et al, 2012 ; Willcox et al, 2012 ; Heslehurst et al, 2013 ; Knight-Agarwal et al, 2014 ). Interviews were transcribed verbatim and data were analysed thematically; data from the practice diaries were also analysed thematically.

Philosophical underpinning

This study drew on the pr inciples of social constructionism, a philosophy rooted in the social sciences ( Franklin, 1995 ; Burr, 2015 ; Gergen, 2015 ) that is concerned with the ‘language’ and ‘traditions’ of particular groups ( Gergen, 2015 ). Gergen (2015) refers to the ‘game of words’ and suggests that individual groups may all have constructed particular ‘languages’ and ‘behaviours’, and that the meanings of particular words, phrases and actions may differ in different contexts.

Community midwives in South East Scotland are currently ‘based’ in community settings such as GP surgeries and work both independently and remotely from their hospital-based colleagues. They are often the first point of contact for pregnant women ( NHS Quality Improvement Scotland, 2009 ) in Scotland. It is possible that the consequences of this contextual situation where midwives meet women (often for the first time) and aim to concurrently raise sensitive topics and build supportive relationships may be influential in ‘constructing’ language, practices and behaviours that are unique to community practice and influence how professional conversations develop.

Data analysis

An iterative, seven-stage, step-wise approach was developed ( Table 6 ) ( Miles and Huberman, 1994 ; Gibbs, 2007 ; Harding, 2019 ) and data were interrogated, observing for language, traditions and behaviours of midwives as they interacted within their contextual situation ( Gergen, 2015 ). Interrogation of the data continued at each stage of analysis, with observance for unique linguistical and behavioural nuances continuing. Three overarching themes were developed that encompassed key elements of community practice: the contextual situation of practice, constructing relationships with women and the midwife as a public health agent. The data were further analysed and interpreted and sub-themes developed. A summary of these themes and sub-themes can be seen in Table 7 .

Description of area of practice from Scottish Index of Multiple Deprivation (2018)

Findings: constructing relationships with women

Midwives appeared to prioritise the development of respectful and trusting relationships with women and their families during the pregnancy and understand that they had a professional duty to provide evidence-based advice:

‘…You have to get their trust, you have to get their respect as the midwife for that first appointment and you can engage them and they feel and they think you're of value to them as the midwife and that you're educating them and can give them good advice then you're on their side and they'll come back but if you alienate them at that first appointment and start lecturing them and you know…’ (Frances)

However, the subject of weight and weight management appeared to be viewed as a sensitive one that some midwives veered away from for fear of causing offence to women in their care:

‘…I don't actually have a problem with it [discussing having a raised BMI] because I really feel it's a benefit to them that eh … I think it's a benefit to them in that we … because they are at huge risk, that we can hopefully reduce this risk and that we'll pick up problems sooner rather than later…’ (Anna)

‘…at that initial meeting, you are trying to build up that relationship and the last thing you want to do is annoy…’ (Catriona)

The above two quotes demonstrate that despite understanding the potential consequences of being obese, tensions arise for professionals that appear to inhibit meaningful discussion.

Midwives turned to relevant clinical protocols as a ‘way in’ to the conversation and viewed these documents as ‘giving permission’ for them to raise sensitive topics of which living with a BMI≥30 kg/m² is one:

‘…Yes, yes, because it's not me having a pop at you. It's me being told this is what I have to do by clinical guidance from the hospital. Then makes them go [name]'s not having a pop at me and saying I'm fat…’ (Elaine)

Using protocols in this way, however, inhibits the development of autonomous practice and suggests that some midwives may be missing opportunities to explore lifestyle choices such as diet, weight maintenance and physical activity, all of which have been shown to improve outcomes for pregnancy, labour and birth ( Jewell et al, 2014 ; McGiveron et al, 2014; Haby et al, 2015 ). This calls into question their adherence to the code ( Nursing and Midwifery Council [NMC], 2018 ) and their understanding of how to practice autonomously, an element of which is to provide current evidence-based advice to women without first requiring permission to do so.

Despite the existence of clinical protocols, some midwives did appear to develop communication strategies of their own volition, attempting to raise conversations with women about living with obesity:

‘…I just think that's just years and I think realising what works and what doesn't and also the feedback from my women because I have been here for 13 years, a lot of them say that they like the ‘spade-like’ approach that I have…’ (Elaine)

However, when probed, this midwife did acknowledge that her communication style was based on her own assessment of her skills and not on evidence-based models. This calls in to questions whether midwives measure their self-assessment against professional standards and may suggest a lack of self-awareness with respect to their professional communication styles.

Information exchange between midwife and woman

When probed more closely, participants acknowledged that they had learnt ‘communication’ informally and ‘on the job’ as a student and during many years of practice rather than as a result of receiving formal education:

‘…I think as a student you've got such a fantastic position, you're in such a fantastic position because you get to work with different midwives and see how they do things and I was forever, “Oh, I like the way she said that”, “Oh, I like the way she did that”, you know. So you pick up a lot of positive ways of saying things through watching other people and then I think when you qualify you just have to learn what, what your style is but it takes time to do that you know…’ (Linda)

Although there was awareness that individuals used this ‘on-the-job’ approach to develop their communication skills, there appeared to be a lack of insight about various communications styles that are available ( Silverman et al, 2013 ). It is possible that being exposed to educational opportunities that facilitate exploration of these styles may support practice.

‘The protocols say’ the risk to professional judgement

Institutional questionnaires used to record information in the maternity notes appeared to not only provide permission for midwives to raise the topic of obesity, but also to devolve the responsibility for doing so to the institution:

‘…So I say “the computer's just worked out your BMI, were you aware of your BMI, do you know what BMI is?” and just, you know, explore it with them. And they just go, “Well I know I'm a wee bit heavier” or sometimes if it's a second-time mum they'll go, “Well I'm definitely a wee bit heavier than I was in my first pregnancy” … ’ (Frances)

In addition, there were several instances where midwives appeared to utilise these questionnaire as a ‘tick-box’ list, rather than personalising and structuring the appointments ensuring the inclusion of woman-specific topics:

‘…Basically, I wait to the point in the notes where I'm almost given licence to talk about it because Trak [computer programme] has told me to talk about it. And the same goes for any other kind of awkward questions, like you know, the questions about domestic violence, questions about previous drug use or depression or whatever. I ask the question when it actually occurs in the notes…’ (Kat)

This use of questionnaires appears to perpetuate the lack of autonomous practice and facilitates midwives to ‘hide’ behind the questions, suggesting a ‘self-protection’ strategy, although it was not clear what they were protecting themselves from.

The importance of the midwife-woman relationship (Foster and Hirst, 2014; Jomeen, 2017 ; McParlin et al, 2017 ) is advocated in the professional driver documents as being a positive force in a woman's life ( Pathways for Maternity Care, 2009 ; NHS, 2017 ; Chief Nursing Officers of England, Northern Ireland, Scotland and Wales, 2020). Participants valued the women-midwife relationship, appearing to prioritise it over delivering evidence-based advice that may optimise the health of mother and baby. These findings mirror other international findings ( Heslehurst et al, 2013 ; Knight-Agarwal et al, 2014 ; Pan et al, 2015 ; McParlin et al, 2017 ). This apparent lack of professional awareness calls into question their compliance with the NMC (2018) code of conduct and their understanding of ‘practise effectively’ concerning their responsibility to deliver evidence-based information.

Participants appeared to pride themselves on being ‘good’ communicators; however, they were unable to articulate how they arrived at this conclusion, calling into question their professional knowledge about the communication of sensitive issues. They appeared to judge for themselves, what was ‘good’ or ‘sub-optimal’ practice prior to developing or constructing their own practice rather than underpinning it with evidence. This suggests that their professional communication skills may not be as effective as they believe. Silverman et al (2013) caution that ‘experience alone can be a poor teacher’ and state that communication education should be as robust as that of any other skill. Provision of targeted education with respect to communication may therefore, strengthen midwifery practice.

Midwives viewed the clinical protocols as providing ‘permission’ to raise ‘sensitive’ topics of which obesity is one. This appeared to allow them to ‘distance’ themselves from the somewhat prescriptive obesity protocol used in the health board (Denison et al, 2016) and the investigations that are advised for obese pregnant women, explaining that these were compulsory requirements rather than being optional. This risks midwifery practice being viewed from a medical perspective. Phrases such as, ‘I have to [instigate various additional investigations] reflect this more medicalised approach, suggesting that women are being failed by midwives who do not always seek informed consent which is a central tenet of healthcare ( Chan et al, 2017 ). This again calls into question their professionalism and adherence to the code ( NMC, 2018 ).

The institutional questionnaires that are utilised to record information during antenatal appointments appear to provide a ‘checklist’ type approach to the antenatal interview. Midwives appear to find this helpful, however, this means that all women are asked the same questions irrespective of their individual needs or wishes. This risks conflating the ‘woman-centred’ appointment to a list of routinised questions that may devalue professional autonomy and lead to ineffective individualised decisions being made ( Frain, 2018 ).

Providing educational opportunities for midwives at both undergraduate and postgraduate levels that facilitate exploration around communication and consultation skills may strengthen and support practice. Recipients of such education may, as a result, develop confidence in raising sensitive issues while concurrently developing effective partnerships with women.

Limitations

This was an exploratory study with 13 practicing community midwife participants. While offering insight into how midwives practice with respect to providing care to obese pregnant women in one geographical location, it is unclear if these findings can be generalised to other areas of practice or geographical areas. All data were collected by one researcher. The views of women were not captured during this study but this was a conscious decision because the research aimed to explore influencing factors for midwifery practice during antenatal appointments. Comparisons between length of experience and practice style was not investigated, doing so may have elicited further insights into practice. While midwives appeared anxious about offending women, none of them raised concerns about having complaints raised against them. Further enquiry about this issue may also have yielded deeper insights.

The observational way in which midwives reported that they learn communication skills may be a weakness in professional practice. Midwives reported that they had not received any formal education around communication although this was not verified with Higher Education Institutions (HEIs). Midwives are failing women by omitting appropriate evidence-based information during appointments for some women. This in turn could risk the good health of mother and baby.

Designing curricula, ensuring that targeted educational components, with respect to communication and consultation skills are included may strengthen midwifery practice and empower midwives to raise and discuss the topic of obesity. However, more work is required to ascertain if this educational need is UK-wide. In addition, the provision of postgraduate educational opportunities by HEIs that will facilitate midwives to develop their communication and consultations skills may be beneficial for those already in practice. Midwives may not be aware that they are contravening the code ( NMC, 2018 ) by not raising and maintaining dialogue about obesity with pregnant women.

• Rates of obese pregnant women presenting for maternity care continue to rise
• How midwives raise and maintain dialogue with women about health risks is poorly understood
• Evidence suggests that modifying diet, increasing physical activity and maintaining weight in pregnancy can improve maternity outcomes
• Research was undertaken to gain deeper understanding about how community midwives approach this element of care
• Findings suggest that providing educational opportunities at both undergraduate and postgraduate levels with respect to communication and consultation may strengthen practice and empower midwives to engage in conversations about perceived sensitive topics

CPD reflective questions

• Consider your understanding of the differences between communication and consultation. Are you clear on the definition of either/both?
• What consultation models, if any, are you aware of that could be used to underpin and influence your practice to facilitate effective discussions surrounding obesity in pregnancy?
• How much consideration have you given to the code ( NMC, 2018 ) and how you may be contravening it by avoiding discussion around this topic?
• Considering your earlier answer, how will you adjust your practice for future consultations with obese pregnant women?

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• HHS Author Manuscripts

BMI is Just a Number: Conflating Riskiness and Unhealthiness in Discourse on Body Size

Iliya gutin.

1 Department of Sociology, University of North Carolina at Chapel Hill

2 Carolina Population Center

Despite the ubiquity of the body mass index (BMI) in discourse on health, there is ambiguity in its use as a biomarker of current abnormality versus future risk. This distinction is consequential for knowledge of the relationship between body size and health, as well as for individuals deemed to have abnormal and “unhealthy” bodies. Consequently, the purposes of this review are threefold. The first is to differentiate this ‘biomarker’ perspective from extant critiques of BMI as a proxy for health behaviors or as the defining characteristic of obesity as a disease . The second is to highlight the shift towards treating BMI as a measure of attained unhealthiness, rather than a probabilistic indicator of risk. Finally, rather than call for the abolition of BMI, this paper argues that its continued use as ‘just a number’ is in keeping with the push for weight neutrality in research and practice. The review concludes by demonstrating how the riskiness and unhealthiness of body size is conflated in public health messaging on COVID-19. BMI is a marker of risk, but its use as a surrogate for COVID-19 severity equates body size with health, shaping beliefs about vulnerability and personal responsibility amid an ongoing pandemic.

Introduction

Over the past 50 years, body size has become a focal dimension of how practitioners, researchers, and the public evaluate health ( Fletcher 2014 ; Jutel 2001 ; Oliver 2006 ; Saguy 2012 ). The ability to quickly and easily measure body size using the Body Mass Index (BMI; body weight [kg] / height-squared [m 2 ]), and then map these values onto substantive categories like “overweight” and “obesity,” has made body size pervasive in scientific and lay narratives about the healthiness of a given person or population ( Nicholls 2013 ). Indeed, BMI is a progenitor to the development and growth of biosocial data in recent decades, wherein the collection of biomarker data allows sociologists and other researchers to obtain more direct and objective evidence of how social forces and environments contribute to bodily ‘wear and tear’ ( Harris and Schorpp 2018 ; McDade 2008 ).

Indeed, the prevalence of BMI in research and clinical settings in the 21 st century is ever-increasing ( Gutin 2018 ), attributable to its convenience of measurement ( Burkhauser and Cawley 2008 ), ease of interpretation ( Fletcher 2014 ), and an innate appeal to societal beliefs about how an individual’s appearance reflects underlying health ( Jutel and Buetow 2007 ). BMI is an established and conventional measure; it is sufficiently widespread and accepted that its orthodoxy is largely unchallenged. Certainly, many have noted that BMI last many limitations as a measure of the harmful adiposity and unhealthy body composition that researchers want to identify ( Rothman 2008 ). However, convenience outranks construct validity, as the tradeoff between expedience and accuracy is one many are willing to make in using BMI to study population health ( Nicholls 2013 ).

Consequently, the purpose of this review is not to relitigate the measurement issues surrounding BMI, or call for replacement ( Kragelund and Omland 2005 ). Rather, the goal is to discuss ontological issues in how BMI is thought about and used, wherein social theory is instructive. Undoubtedly, better measurement is the most direct route by which to improve research; however, a measurement-centered approach is untenable given methodological constraints in extant research and practice, where BMI will continue to be the primary diagnostic tool ( Hu 2008 ). Nor does better measurement – ranging from waist and hip circumference to bioelectrical impedance – provide a full explanation of how a person’s physical appearance or body size and shape is associated with their health and wellbeing ( Yates-Doerr 2013 ). BMI is an imperfect biomarker, but it is also inescapable, such that researchers are better served by improving understanding of how to incorporate and interpret BMI as the measure that is available rather than the measure that is desired . In turn, this review makes a case for better theory , emphasizing that measures of risk are not interchangeable with measures of health, especially when both biophysiological and psychosocial mechanisms are at play – as is true of BMI.

Pursuant of this goal, the paper begins by differentiating the ‘biomarker’ perspective from extant critiques of BMI as a proxy for individuals’ health behaviors or as the defining characteristic of obesity as a disease. Next, the paper demonstrates the shift toward treating BMI as a measure of attained unhealthiness, rather than a probabilistic indicator of risk. Finally, rather than call for the abolition of BMI, the paper advances a novel argument contending that its continued use as ‘just a number’ – with health-relevant attributes – is in keeping with the push for weight neutrality in research and practice. The review closes with an applied example, showing how the riskiness and unhealthiness associated with body size is conflated in public health messaging on COVID-19. BMI is labeled as a risk factor but in practice it is deployed as a surrogate measure of COVID-19 severity, shaping knowledge about who is most vulnerable to and responsible for severe illness in contemporary society.

Not a Behavior, Not Yet a Disease

This review makes a case for a more neutral framing and use of BMI, yet that this is far from the status quo in extant research. Behavioral/lifestyle and disease frameworks have served as the dominant perspective on body size, representing opposite ends of the pathogenic process by which individuals’ actions and exposures are manifest in diagnosable health outcomes ( Kelly and Russo 2018 ). They reflect disciplinary distinctions in how BMI is conceptualized, and the epistemological concerns of those studying health. As will be shown, biomarkers do not represent actions or decisions, nor do they constitute definitive diagnoses. They instead represent a more ambiguous middle ground, linking behaviors to diseases – or actions to outcomes – via a process of mediation often couched in the language of ‘mechanisms’ and ‘pathways’ ( Loucks et al. 2008 ). Beyond their mechanistic function, this review argues that a biomarker perspective allows for a form of ‘neutrality’ that circumvents the negative assumptions that arise from treating BMI as a proxy for harmful behaviors, or as a criterion by which individuals are labeled as ‘diseased.’

Health Behavior or Lifestyle

Treating BMI as a behavioral concern – commensurate with smoking, alcohol consumption, drug use, and other “actuarial,” health-relevant factors ( Jones and Oppenheimer 2017 ) – is exemplary of the contemporary public health paradigm, which seeks to identify and intervene on the harmful behaviors accounting for the majority of morbidity and mortality ( Dew 2012 ). Surveillance, or “surveillance medicine” ( Armstrong 1995 ), is central to this framing; the individual- and population-level tracking of health behaviors becomes the primary way to minimize the uncertainty of risk and poor health in the modern world. In turn, the ongoing surveillance of BMI over past decades has made obesity inseparable from the label of “epidemic” ( Fletcher 2014 ), emphasizing the prevalence of this epidemiologic perspective.

Crucially, this framework invokes BMI as a ‘summary’ measure of various changes in behaviors and lifestyles over past decades. The transition towards more sedentary labor and leisure activities and increasingly calorie-rich environment are hypothesized as key obesogenic lifestyle-relevant mechanisms afforded by modernization (Phillipson and Posner 2003 ). Thus, as public health and medical advances have largely eliminated many communicable diseases and other acute threats to health in many societies – and as other harmful behaviors like smoking continue to decline – obesity has been singled out as the primary lifestyle factor underlying future chronic morbidity and premature mortality ( Stokes and Preston 2016 ).

However, using BMI as a measure of behaviors and lifestyles is laden with assumptions and stereotypes about individuals’ bodies as the literal embodiment of hedonism (Murray 2012; Saguy 2012 ). Disease prevention in public health is predicated on a shift towards the adoption and maintenance of healthy lifestyles, with weight ‘control’ as an illustrative case. Though well-intentioned, this framework advances the idea of people striving to be good “biocitizens,” who adhere to certain health standards that reflect their value as contributing members of society ( Dew 2012 ). This reflexive and recursive link between health ideals and social ideals is unavoidable in the contemporary world, where societal concerns about reducing risk are located at the individual level, as individuals are held ‘responsible’ for minimizing their own risk and, by extension, risk in the population as a whole ( Beck 1992 ; Lupton 2006 ). Health ideals and social ideals are thus comingled and co-constructed, influencing how individuals are judged as a function of whether their bodies are thought to reflect good decisions ( Fox 2012 ).

Indeed, the conflation of epidemiologic and social norms about healthiness has made thinness a marker of one’s social status and morality. Managing one’s health – by means of having a thin, medically ‘appropriate’ body – is a hallmark of the “good” biocitizen ( Greenhalgh 2015 ). Overweight or obesity, as a deviation from a “healthy” and “normal” BMI, represents a flawed identity premised on the inability for self-discipline (Ciciurkaite and Perry 2017; Oliver 2006 ; Shugart 2016 ). Thus, the linkage of body size to other lifestyle attributes like diet and exercise leads to the harmful and unfounded assumption that individuals with overweight and obesity are slothful, greedy, and selfish (Murray 2012; Saguy 2012 ; Shugart 2016 ), despite many adults with “unhealthy” BMIs engaging in the kinds of healthful behaviors that are valued under this paradigm of good biocitizenship ( Bombak et al. 2019 ; Greenhalgh 2015 ; Tylka et al. 2014 ).

While the behavioral framework reflects the surveillance imperatives in epidemiology and public health, conceptualizing obesity as a disease reflects the drive for unambiguous assessments of health in the context of clinical practice. Contemporary biomedicine is premised on the mitigation of uncertainty in the understanding of health and, more importantly, threats to health. The biomedical model makes disease its focus, creating a standard by which normality and health are defined in relation to disease ( Fox 2012 ), with the disease label delineating good and bad health ( Jutel 2019 ). Describing obesity as a disease therefore reduces uncertainty in how much clinical significance health professionals, researchers, and laypersons accord to BMI.

On the one hand, the disease label is a nosologic necessity in the context of medical practice, wherein the nosology – classification and naming – of various conditions and ailments is of the utmost concern ( Jutel 2014 ). Disease categories serve as the gateway for legitimization, treatment, and reimbursement, which permits body size to be worthy of medical attention ( Jutel 2014 ). The disease label opens avenues of treatment that might otherwise be closed due to an individual-level behavioral/lifestyle perspective ( Rosenberg 2002 ). Pragmatism is a common justification for this framing, recognizing that medical care is as much a function of what is considered legitimate and billable, as it is a function of what individual practitioners and patients believe to be important. Many governing medical bodies are explicit about adopting the practice of labeling obesity as a disease in an effort to legitimate obesity as a serious and diagnosable health condition within the eyes of medical practitioners, their patients, and the population at large ( Allison et al. 2008 ; Bray et al. 2017 ; Kyle et al. 2016 ).

On the other hand, the disease label is potentially unjustified and harmful in perpetuating weight-related stigmatization and bias ( Kyle et al. 2016 ; Sharma and Campbell-Scherer 2017 ; Tomiyama et al. 2018 ). While not challenging the notion that excess adiposity may be harmful, labeling individuals with obesity as diseased on basis of BMI does not reflect practitioners’ knowledge of a person’s overall health and the extent to which it is directly affected by their weight ( Sharma and Campbell-Scherer 2017 ). This conflation of measurement with a formal disease label undermines the conceptualization of disease as a distinct state of physiological impairment typically used in medical practice ( Charrow and Yerramilli 2018 )

Diagnoses and disease labels are descriptive but disruptive social categories, shaping how individuals view themselves and others, and having consequences that extend far beyond the purview of biomedicine ( Jutel 2019 ). Though the disease label reflects an earnest effort to counteract narratives of personal culpability ( Jastreboff et al. 2019 ), this decision cannot be disentangled from the consequences of labeling ‘abnormal’ bodies as diseased and unhealthy in a society where body size is seen as the product of harmful individual choices ( Brownell et al. 2010 ). Just as the disease label legitimizes obesity in healthcare, it legitimizes stigma towards individuals whose bodies do not conform to the social norms about who is healthy and ‘good.’ Thus, there is a certain injustice in labeling a substantial proportion of the population as “diseased” on the basis of their BMI, when there is not only a lack of knowledge as to whether a person is truly in poor health, but also a lack of safe and sustainable means to make them “well” and “normal” ( Greenhalgh 2015 ).

A Marker of Risk, Not Current Health

The behavioral/lifestyle and disease frameworks are subject to a similar epistemic fallacy surrounding BMI as a measure of individuals’ current health, broadly defined. Epistemologically sound measures should reflect the underlying ontological domains of interest; yet, BMI is rooted in an ontology of probability and risk ( Nuttal 2015 ), or what a person’s body size suggests about future , rather than current, health. As this review contends, recognizing BMI as an imperfect biomarker speaks to this ontology in allowing for uncertainty in how and why body size is associated with specific health outcomes.

Biomarker for Abnormality

Though a universal definition of biomarkers is lacking, the National Institutes of Health’s Biomarkers Definitions Working Group (BDWG) identifies biomarkers as “objectively measured and evaluated” indicators of “normal biological processes [and] pathogenic processes,” encompassing numerous measures in recent years ( BDWG 2001 ). Critically, the fact that biomarkers often stem from behaviors does not make them substitute measures. Blood pressure, lipids, and glucose are associated with similar behaviors as BMI, but they are not proxies for health lifestyles; indeed, these biomarkers are granted more latitude in what they suggest about a person’s choices – or lack thereof – to the extent elevated blood pressure or inflammatory markers reflect elevated stress from external stimuli ( Harris and Schorpp 2018 ).

From a disease perspective, biomarkers represent pathways rather than definitive, diagnosable clinical endpoints ( BDWG 2001 ). They are integral to disease diagnosis ( Timmermans and Haas 2008 ), but a transition to an ‘abnormal’ value for a biomarker does not demarcate a latent transition from disease-free to diseased. Admittedly, diabetes and hypertension are near-exclusively defined on the basis of blood glucose and blood pressure, respectively, exceeding a predetermined threshold. However, BMI is distinctive given the pragmatic, rather than epistemic, choice to define obesity as a disease based on BMI ( Bray et al. 2017 ). No biomarker is a perfect measure of the underlying biological or pathogenic process it represents ( Loucks et al. 2008 ); indeed, very few biomarkers meet the criteria for being “surrogate endpoints,” in fully mediating the relationship between an exposure and a clinical endpoint ( BDWG 2001 ; Loucks et al. 2008 ). But it is important to recognize that biomarkers exist on a spectrum of surrogacy, which should be acknowledged in their use. There is a tight conceptual linkage between the pathogenesis of diabetes and hypertension and their underlying biomarkers; by contrast, obesity is described as unhealthy or excess adiposity contributing to cardiometabolic abnormality, neither of which is perfectly measured or mediated by BMI. Per this logic, the proximity of a biomarker to its clinical endpoints (i.e., its degree of surrogacy), should dictate its use as a marker of risk as compared to a marker of current health.

Yet, in practice, BMI has effectively become a surrogate marker of present physiological abnormality or ‘unhealthiness’, as a nonspecific clinical endpoint ( Jutel 2014 ). The notion of risk – and uncertainty – is replaced with a more definitive assessment of individuals’ proto-illness/disease status based on measured risk ( Aronowitz 2009 ; Gillespie 2012 ; Gillespie 2015 ; Jauho 2019 ). One’s health is assumed to be impaired when exceeding a “normal” BMI range, as abnormality constitutes evidence of poor health, or the point at which BMI represents “a social condition of compromised health status” ( Gillespie 2012 : 195). There is strategic value to using BMI as a surrogate marker. Contemporary health is shaped by standards for how much significance is afforded to measures in terms of their ‘biomedical’ worth as evidence ( Timmermans and Berg 2003 ). Yet, evidence is judged on its ability to provide “transparent, objective, and irrefutable information about the body” ( Jutel 2014 : 124); consequently, priority is accorded to measures that are quantifiable and standardized, and largely free of any uncertainty or ambiguity in what they imply about individuals’ health.

In the case of BMI, researchers and practitioners strive to impose certainty – and assume surrogacy – to compensate for the fact that BMI does not provide “transparent, objective, and irrefutable information” to the extent that these attributes are desired in biomarkers. In theory, biomarkers convey risk about the development of a given condition or disease. They are relationally defined, as a function of future morbidity, to the extent that there is confidence in their ability to provide evidence of a biophysiological mechanism. The central issue in using BMI is an inability to clearly identify poor health beyond tautologically describing BMI as a sign of impairment ( Sharma and Campbell-Scherer 2017 ). Yet this has not curbed the reification of the association between body size and health in using BMI categories to sort individuals by their degree of healthiness, regardless of how they fare otherwise ( Jutel 2011 ).

The conflation of future risk and present health with respect to BMI stands in contrast to the numerous other ‘risk’ factors strongly associated with varying degrees of elevated risk for morbidity and mortality – such as being male, over the age of 40, riding a motorcycle without a helmet, and not getting enough sleep ( Allison et al. 2008 ). None of these risk factors are used as direct measures of health; nor are they used as surrogate markers to represent a pathology of poor health, even though many of these factors are more strongly associated with increased morbidity and mortality than BMI. For instance, smoking is so strongly associated with lung cancer that one could use this same crude logic to argue that smoking is a surrogate measure for the disease ( Aronson 2005 ). Yet smoking remains a risk factor while BMI is used as a surrogate for obesity as a diseased state ( Jutel 2006 ). Certainly, the preference is that individuals are free of these external risks, especially from modifiable factors, but such preferences for living in a ‘risk-free’ society should not serve as the basis of perceptions of, beliefs about, and interactions with individuals as being unhealthy and deviant, to the extent this is the true of BMI.

Just a Number

While BMI is a far from a perfect biomarker, from the perspective of both measurement and theory, one cannot ignore the wealth of evidence documenting strong associations between BMI and multiple dimensions of health. Yet, in explaining these associations, one also cannot ignore the discrepancy between the literal definition of BMI, based on height and weight, and the extent to which BMI is used as a surrogate for myriad health processes. These associations have become interchangeable with the values and categories of BMI that are used by researchers, practitioners, and individuals, ignoring that BMI is ‘just a number’ ascribed with meaning in different contexts that suit various stakeholders’ needs.

Namely, this ‘just a number’ framework is intended to downplay the power and specificity of meaning afforded to measures like BMI, in recognizing that conceptualizations and definitions of health are often a reflection of changing norms about what constitutes objective evidence, rather than changes in the objective reality of health ( Timmermans and Berg 2003 ). The unchanging, objective reality of BMI is that – above all else – it is always the same number, calculated from the same equation, while the meanings it is imbued with are constantly in flux as a function of time, place, and the needs of individuals and institutions using it in various contexts. Its meaning and usage may be perfectly appropriate in one scenario, but not in another, despite the number remaining unchanged. Moreover, knowledge of how BMI is used in one context can be informative for the practice of its use in another, but these opportunities to broaden its interpretation are only possible when BMI is viewed as a number rather than as a surrogate for a specific domain of health.

The quantifiability of BMI has proved especially valuable given modern standards of objective and standardizable evidence; yet, numbers are rarely used in a ‘neutral’ manner. There is a societal propensity to “ordinalize” and “categorize” knowledge into heuristic rankings and groupings ( Bowker and Star 1999 ; Fourcade 2016 ), using values like BMI to compare individuals and establish hierarchies of on the basis of weight and assumptions about health. Thus, in an effort to subvert the societal and scientific imperative to label and sort BMI, a biomarker-informed perspective on BMI as ‘just a number’ is arguably the most appropriate way of framing body size, recognizing that biomarkers can convey risk without being declarative about health.

In and of itself, BMI is neutral on the subject of an individual’s current health or the degree to which it conforms to ‘normal’ physiological functioning; imbuing the measure with additional assumptions creates linkages to abstract ideas about unhealthiness and abnormality. While there is guidance and helpful theory on how to use biomarkers ( Harris and Schorpp 2018 ; McDade 2008 ; Timmermans and Haas 2008 ), there is less consideration of why some biomarkers become imbued with additional meanings outside their function as biophysiological mechanisms connecting exposures to outcomes. This mediational framework is contingent on biomarkers being reasonable measures of the physiological processes taking place, maximizing how much of the effect of the exposure on the outcome is accounted by the biomarker ( Loucks et al. 2008 ). BMI, as ‘just a number,’ does not necessarily satisfy this framework, and additional explanations are sought out to prevent it from collapsing.

Consequently, researchers often only care about BMI inasmuch as it purports to convey information about other aspects of health, like the types of behaviors a person engages in or their level of physiological impairment, to the extent that having a scale or measurement like BMI facilitates comparisons and rankings ( Bowker and Star 1999 ; Fourcade 2016 ; Jutel 2006 ). These behavioral and disease frameworks are inherently attractive as they convey definitive information about what a person is doing or how they are, at present, rather than describing health in probabilistic terms (e.g., an individual has a BMI in a range associated with a 20–30% higher risk of mortality, on average, compared to individuals with a BMI in a “normal” or “healthy” range). Avoiding, rather than acknowledging, uncertainty in describing body size pushes us to make more definitive pronouncements about individuals’ healthiness based on their BMI. But not all biomarkers allow us to make such clear assessments of underlying health as compared to underlying risk ( Loucks et al. 2008 ). Health is a multifaceted, systems-level construct (Harris 2010; McDade 2008 ); individual biomarkers help identify its discrete components, but they do not serve as the basis for comprehensive judgements. Moreover, this perspective neglects the possibility that these measures associated with health through diverse mechanisms, some of which are primarily psychosocial rather than biophysiological. Thus, the challenge in working with BMI is preserving its neutrality by resisting the temptation to situate it in a specific domain of health.

Social Marker of Inequality

Mapping BMI onto tangible domains of health – whether behavior, disease, or as a biomarker of physiologic abnormality – represents the dominant ontological approach. However, there is increasing recognition of an alternative perspective that bridges clinical and epidemiologic research on the limitations of BMI as a health surrogate with a sociological and psychological understanding of body size as an axis of inequality. This ‘weight neutral’ framework does not downplay the importance of studying body size and health; rather, it downplays the need to directly and unambiguously equate body size with health in research, practice medicine, and the conceptualization of overweight and obesity. Body size is acknowledged as a neutral form of human variation ( Saguy 2012 ), whereby BMI reflects both biophysiological and psychosocial mechanisms of risk. To the extent that BMI is surrogate marker of physical appearance, it is simultaneously marker of social abnormality and inequality, representing a distinct process by which individuals’ social interactions and experiences affect their health. Consequently, there is considerable heterogeneity in how and why researchers explain the relationship between BMI and health. In turn, treating BMI as ‘just a number’ provides a broader set of explanations and points of intervention.

Health at Every Size

More than just an abstract concept, weight neutrality is seen as a practical and sustainable alternative to the BMI-centric approaches used in assessing and intervening on individual and population health. The Health at Every Size (HAES) movement promotes weight neutrality in pushing researchers and practitioners to acknowledge the limitations of BMI as a defining characteristic of health, and recognize diversity in weight and health, in considering alternative explanations for why BMI confers higher risk. Promoting good health is a priority, rather than explicitly healthy weight, given that an overly-narrow focus on attaining the latter is often harmful, in and of itself, with respect to poor mental health, disordered eating and exercise, and how these psychosocial aspects of health affect physiological functioning ( Bombak and Monaghan 2017 ; Bombak et al. 2019 ; Tylka et al. 2014 ).

Weight-targeted interventions are often ineffective, emphasizing dietary and exercise regimes for which success is measured solely by weight loss ( Mann et al. 2007 ). Many adults can successfully and sustainably improve many other cardiometabolic indicators that allow for better overall health and longevity ( Bacon and Aphramor 2011 ; Mann et al. 2007 ; Tylka et al. 2014 ). Indeed, discordance between individuals’ having an “unhealthy” BMI despite normal measures of cardiometabolic functioning ( Greenhalgh 2015 ; Saguy 2012 ), and a general sense of robustness ( Monaghan 2007 ), creates additional stress and anxiety surrounding what it means to be ‘healthy.’ In a society where body size is a source of stigma ( Puhl and Heuer 2010 ), a focus on BMI can be limiting and distracting during a medical encounter ( Phelan et al. 2015 ), leading to skewed assessments of individuals’ health which are magnified when they serve as the basis for population-level guidance and policies on what constitutes good health and a healthy body.

Essentializing BMI

Weight neutrality is premised on the fact that conflating BMI, risk, and health is damaging and unjust, in reinforcing stereotypes about individuals who do not adhere to social norms for physical appearance and health, and how the two are equated on the basis of social norms for beauty and fitness ( Jutel and Buetow 2007 ). The notion that phenotypic attributes become imbued with social meaning – and thus become health-relevant traits – is not a novel concept (Link and Phelan 2001). Directly equating body size with race is too strong a comparison, to the extent that race is tied to endemic legacies and systems of oppression ( Phelan and Link 2015 ), but one should not ignore how BMI and race exemplify how one’s phenotype affects health though non-biophysiological pathways.

The issues surrounding race as an essentialized concept provide a clear illustration of how phenotypic traits are conflated with their social consequences ( Frank 2007 ; Gutin 2019 ; Morning 2011 ), wherein race, itself, is assumed to be the innate, causal mechanism underlying poor health. Yet, decades of research prove that the relationship between race and health is attributable to race being a proxy for the many social ills inflicted upon non-White persons via interpersonal and institutional forms of discrimination and disenfranchisement ( Phelan and Link 2015 ). Unfortunately, this message fails to resonate in a society where health is actively used to gauge individuals’ social standing ( Scambler 2009 ); the moral judgment attached to healthiness substantiates the belief that those who are unhealthy are ‘bad’ members of society. This gives rise to a vicious cycle by which the poor health of a marginalized group is used to justify their marginalization, likely leading to worse health in the future.

A comparable process of essentializing BMI has been at work for decades, legitimizing BMI as a surrogate marker of biophysiological health, while ignoring the psychosocial implications of its being a marker of appearance and status. Once again, tautological reasoning is partially to blame; a person becomes unhealthy upon attaining an unhealthy BMI, implying some kind of transition in their latent health. Body size has been problematized and stigmatized as an abnormal form of human variation just as other forms of human variation have been considered ‘undesirable.’ In a highly weight-conscious society where morality is linked to one’s appearance ( Jutel and Buetow 2007 ; Shugart 2016 ), body size represents another form of phenotypic stratification that influences health. Stigma is a fundamental mechanism underlying health disparities ( Hatzenbuehler et al. 2013 ), with body weight being one of the earliest forms of stigma examined by sociologists ( Cahnman 1968 ; Maddox et al. 1968 ). Yet, nearly 50 years on, it continues to be a “socially acceptable form of bias” ( Puhl and Heuer 2010 : 1019), and the association between having an abnormal body and poor health as a given ( Greenhalgh 2015 ).

There is strong evidence to suggest that stigma underlies mechanisms linking obesity to numerous health conditions, independent of the physiological consequences of BMI. Institutional biases in the workplace, educational settings, healthcare, interpersonal relationships, and media lead to worse treatment and fewer rewards for individuals with overweight and obesity ( Puhl and Heuer 2009 ), directly impacting their socioeconomic prospects, quality of life, and health. Moreover, the omnipresent stigmatization of body size is manifest as discrimination, ostracism, harassment towards those with higher BMIs, and the internalization of negative self-imagery among individuals whose bodies do not conform to social and medical ‘norms’ ( Puhl and Heuer 2010 ; Puhl et al. 2008 ). Weight-based stigma is associated with numerous mental and physical health outcomes such as depression, anxiety, psychiatric disorders, impaired cardiovascular health, and many others ( Hatzenbuehler et al. 2009 ; Papadopoulos and Brennan 2015 ; Puhl and Heuer 2010 ; Puhl and Suh 2015 ; Schafer and Ferraro 2011 ). The cumulative impact of these chronic insults is implicated in the association between weight-related stigma and premature mortality ( Sutin et al. 2015 ). In their totality, there is compelling evidence that psychosocial mechanisms constitute some of the primary pathways through which individuals’ body size negatively impacts their health ( Tomiyama et al. 2018 ). Researchers often lack direct measures of weight-based stigma, but this does not excuse ignoring such explanations when interpreting BMI as a ‘surrogate’ measure.

More broadly, BMI is a marker for health in the same way phenotypic attributes like race and gender are determinants of health; they gauge future risk rather than serve as measures of current health. Certainly, BMI is distinctive – and challenging – as it is not exclusively a marker of appearance and has real biophysiological consequences. Studying the relationship between body size and health is important, but there is a need to better acknowledge uncertainty in what BMI serves a marker of . The conceptualization of race in health research continues to serve as a useful parallel; there are legitimate concerns about how race being used to perpetuate biogenetic explanations ( Bliss 2012 ; Shim 2002 ; Smart and Weiner), but recognition that race is a socially-meaningful category is vital for advancing justice and equity ( Borrell et al. 2021 ; Epstein 2008 ). Thus, rather than limit discussion to biophysiological explanations for why BMI is associated with adverse health, the inequality framework allows for a broader set of psychosocial pathways and interventions.

BMI in the Time of COVID-19

The need for conceptual clarity in what BMI measures and how it becomes associated with health is not an abstract concern, given that many clinicians, epidemiologists, and public health officials rely on the measure to make decisions about individual and population health. It is beyond the scope of this review to consider the myriad diagnoses and interventions where more careful use of BMI may alter experts’ interpretations and resulting course of action. However, recent discourse on COVID-19 in the United States serves as an illustrative and timely example of how BMI has effectively become a surrogate for severity of illness. The fact that BMI reflects adiposity and biophysiological abnormality that puts one at elevated risk for severe infection and mortality is undeniable ( Popkin et al. 2020 ); yet, non-biophysiological factors cannot be ignored ( Hill et al. 2021 ; Townsend et al. 2020 ). As with numerous other conditions, in treating BMI as a marker of health, rather than risk, the scope of plausible explanations is unnecessarily limited.

Conflating risk and health in discourse on BMI creates confusion about vulnerability amid existing uncertainty about what it means to be a safe and responsible member of society, especially when this messaging plays a key role in shaping public knowledge of healthiness, morality, and even social status (Monaghan 2021). The risk factors identified by the Centers for Disease Control (CDC) in the United States represent an ontological mélange of health behaviors, biomarkers, and conditions or diseases, such as smoking, high blood pressure, and diabetes ( CDC 2021 ). There is a ‘hierarchy’ of risk, in using these factors to separate individuals who “might be” at risk from those who already “are.” Yet, the only risk factor that appears throughout this hierarchy is BMI – in distinguishing between overweight, obese, and morbidly obese – whereby the graded relationship between body size and COVID severity mirrors the narrative of BMI as a surrogate for latent biophysiological health vis-à-vis cardiometabolic and immune functioning ( Kwok et al. 2020 ). Decisions about how to categorize risk are difficult, reflecting empirical evidence linking BMI to COVID-19 outcomes. However, it is important to recall that this evidence is associational , and far from neutral in reinforcing a clinical, biophysiological conceptualization of increased risk as indicative of worsening health ( Gillespie 2012 ; Jutel 2011 ). In other words, emerging narratives actively contributed to individuals’ current obesity as effectively being a surrogate marker of their future COVID-19 risk, and severity therein, adding to a long list of health conditions for which BMI is used as a proxy.

How risk is conceptualized and conveyed reflects different narratives about vulnerability and underlying mechanisms. Deploying the neutral perspective on BMI as ‘just a number’ can be instructive in communicating that the association between BMI and COVID-19 is a multidimensional process of increased physiological risk compounded by the stigma of medicalized identities, as well as individuals’ social environments and contexts ( Puhl et al. 2020 ). There is concern about BMI being a marker of social status amid the pandemic, reflected health professionals’ assumptions about individuals with obesity and emerging media narratives about who is to blame. These messages reinforce social beliefs about individual responsibility for worse health among people with higher BMIs, which may exacerbate COVID risk by increasing their propensity for unhealthy behaviors or avoiding needed medical care ( Hill et al. 2021 ; Flint 2020 ; Le Brocq et al. 2020 ; Puhl et al. 2020 ; Townsend et al. 2020 ; Wu 2020 ).

Race and ethnicity are also ‘risk factors’ given the disproportionate toll of COVID-19 on minority individuals and communities ( CDC 2020 ). However, it is apparent that greater vulnerability and exposure owing to a multiplicity of sociostructural mechanisms – such as systemic disinvestment, material deprivation, and institutional racism ( Phelan and Link 2015 ) – shapes worse COVID-19 outcomes for non-White persons, rather than race representing a direct causal link between race and severe illness ( Hooper et al. 2020 ; Krieger 2020 ; Yancy 2020 ). Many of these mechanisms of discrimination, bias, and ignorance are well-known explanations linking body size to health, and serve as plausible mediators for the association between BMI and COVID-19 severity ( Hill et al. 2021 ; Townsend et al. 2020 ), independent of biophysiological pathways. Yet these sociostructural explanations are ignored when the neutrality of BMI is confounded by pre-existing assumptions about body size, biocitizenship and healthiness.

Ultimately, the issue of delineating biophysiological from sociostructural risk had direct bearing on determining who was prioritized for COVID-19 vaccinations. Having a BMI greater than 30 made one ‘eligible,’ entirely from a biophysiological risk perspective, in representing a comorbidity ( CDC 2021 ). As noted, these decisions are difficult, and all efforts should be made to prioritize care for those at the highest risk of severe illness, as seen in data on BMI. It is also understandable that definitions of biophysiological risk and unhealthiness err on the side being liberal, with the goal of unintentionally excluding risk factors because the exact mechanisms of action are unclear. Despite the strong, ethical case that a sociostructural definition of risk would allow for a more inclusive definition of eligibility – especially in the case of high-risk minority populations ( Schmidt et al. 2020 ) – the current emphasis on biophysiological risk reflects extant norms about using BMI as a flawed, but acceptable, surrogate of health.

Given limited initial supply, and far greater demand, the debate over who was sufficiently ‘at risk’ to be vaccinated intersects with social norms about personal responsibility and health. BMI categorizes over 40% of the U.S. adult population as obese and ‘unhealthy’ ( Hales et al. 2020 ); it is the most prevalent risk factor and source of eligibility. Certainly, many adults for whom body weight is a health issue stand to benefit, just as they have from the legitimization of obesity as a medical issue in other contexts. Yet, given the inadequacy of BMI as a measure of individual health, it is likely that not all vaccines will go to individuals who are most in need. The implications this may have in perpetuating weight bias and stigmatization are unclear, especially if society’s propensity to assign personal responsibility for higher BMIs is conflated with the misperception that irresponsibility and poor health is being ‘rewarded’ with vaccination. Health is a well-established source of stigma in society ( Scambler 2009 ), and the imprecise use of BMI as a measure underlying health has potential to do harm. Thus, just as there is concern with the long-term consequences of COVID-19 as a disease, it is important to consider the ramifications of how conflating obesity and COVID-19 as intersecting pandemics perpetuates a narrative of individual culpability and social burden that continues after COVID-19 is no longer a threat.

Sociologists, public health scientists, and numerous other researchers engaged in the study of population health are in a difficult position when it comes to understanding body size and health. The work is important and necessary; even if the conceptualization and measurement of an “unhealthy” body size has limitations, one cannot ignore the totality of evidence identifying increased body size as a key factor underlying morbidity and mortality ( Stokes and Preston 2016 ). Yet, studying this relationship taps into an enormity of biophysiological, psychological, social, and cultural mechanisms that are beyond the reach of extant approaches to defining a “healthy” body. BMI is, and will continue to be, the dominant metric, short of unforeseen advances that allow researchers and practitioners to easily measure the many indicators associated with poor cardiometabolic health. Given the unlikelihood of this occurring, the key motivation for this review asks how research can better use BMI to improve knowledge of the relationship between body size and health.

Drawing on contemporary critiques of BMI, obesity, and the dominant ontological frameworks used to describe health, this paper highlights how pre-existing assumptions and beliefs bias interpretation of BMI. The health behavior/lifestyle and disease frameworks are consequential, in assigning specific meaning to BMI as a measure of what one does with their body to maintain their health, or how their body is with respect to latent health ( Greenhalgh 2015 ). Situated between the two, biomarkers are perhaps less definitive in suggesting a mediated process by which individuals may be at future risk for poor health without necessarily being unhealthy at present ( Loucks et al. 2008 ). Yet, in practice, treating biomarkers as “surrogate markers” undermines this emphasis on risk, to the extent that biomarkers are substituted for the future concerns they may be associated with ( Jutel 2014 ). Consequently, the risk attached to BMI is a surrogate for current biophysiological abnormality, despite the epistemological inadequacy of equating BMI with “obesity” as a state of health – or whichever clinical endpoint BMI is alleged to represent.

This does not mean that a biomarker approach to conceptualizing BMI is flawed; rather, it is limited. All health measures are ‘just numbers’ until they are imbued with clinical meaning. BMI is laden with many assumptions and beliefs, many extending beyond the realm of biomedicine. It is a valid measure of body size that captures individual’s physical characteristics, and how these qualities come to be associated with health through numerous mechanisms – both biophysiological and psychosocial. This holistic interpretation of BMI not only advances weight neutrality in research ( Gutin 2018 ), but also facilitates the discussion of body size as a source of stigma and axis of inequality – and BMI as a measure of both – in weight-conscious societies ( Puhl and Heuer 2010 ; Saguy 2012 ).

The ubiquity of BMI is attributable to its simplicity as a measure of health; yet, this understates its complexity as a surrogate for numerous biological and social processes. This complexity critical for sustaining a more holistic understanding of the relationship between body size and health. In turn, a balanced framing of BMI in research – accepting of uncertainty and risk, and less declarative about normality and health – can inform broader social norms about diversity in bodies and wellbeing, and how individuals with ‘deviant’ bodies are perceived and treated. The present confusion surrounding BMI as surrogate marker of COVID-19 severity highlights the consequences of how inadequacies in conceptualization create difficulty in using health measures to make difficult policy decisions. Conflating riskiness and unhealthiness – and how this equivalence translates into ideas about who deserves help, or who is responsible for creating a social and health burden – is not an abstract concern. The conceptualization and understanding of BMI has tangible implications for prioritizing care, evaluating the appropriate level of concern, and passing judgment on who is considered to be a contributing member of society.

Acknowledgements

I would like to thank the editors and two anonymous reviewers for their excellent feedback on this manuscript. This work was supported by the Population Research Training grant (T32 HD007168) and the Population Research Infrastructure Program (P2C HD050924) awarded to the Carolina Population Center at The University of North Carolina at Chapel Hill, as well as the Jessie Ball duPont Dissertation Completion Fellowship at The University of North Carolina at Chapel Hill.

• Allison DB, Downey M, Atkinson RL, Billington CJ, Bray GA, Eckel RH, Finkelstein EA, Jensen MD & Tremblay A (2008). Obesity as a disease: A white paper on evidence and arguments commissioned by the Council of the Obesity Society . Obesity , 16 ( 6 ), 1161–1177. [ PubMed ] [ Google Scholar ]
• Armstrong D (1995). The rise of surveillance medicine . Sociology of Health & Illness , 17 ( 3 ), 393–404. [ Google Scholar ]
• Aronowitz RA (2009). The converged experience of risk and disease . The Milbank Quarterly , 87 ( 2 ), 417–442. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Aronson JK (2005). Biomarkers and surrogate endpoints . British Journal of Clinical Pharmacology , 59 ( 5 ), 491–494. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Bacon L, & Aphramor L (2011). Weight science: Evaluating the evidence for a paradigm shift . Nutrition Journal , 10 ( 9 ). [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Beck U (1992). Risk society: Towards a new modernity . (Ritter M, Trans.). Sage Publications. [ Google Scholar ]
• Biomarkers Definitions Working Group. (2001). Biomarkers and surrogate endpoints: Preferred definitions and conceptual framework . Clinical Pharmacology & Therapeutics , 69 ( 3 ), 89–95. [ PubMed ] [ Google Scholar ]
• Bliss C (2012). Race decoded: The genomic fight for social justice . Stanford University Press. [ Google Scholar ]
• Bombak AE, & Monaghan LF (2017). Obesity, bodily change and health identities: A qualitative study of Canadian women . Sociology of Health & Illness , 39 ( 6 ), 923–940. [ PubMed ] [ Google Scholar ]
• Bombak A, Monaghan LF, & Rich E (2019). Dietary approaches to weight-loss, Health At Every Size® and beyond: Rethinking the war on obesity . Social Theory & Health , 17 ( 1 ), 89–108. [ Google Scholar ]
• Borrell LN, Elhawary JR, Fuentes-Afflick E, Witonsky J, Bhakta N, Wu AHB, Bibbins-Domingo K, Rodríguez-Santana JR, Lenoir MA, Gavin III JR, Kittles RA, Zaitlen NA, Wilkes DS, Powe NR, Ziv E, & Burchard EG (2021). Race and Genetic Ancestry in Medicine – A Time for Reckoning with Racism . The New England Journal of Medicine , 384 ( 5 ), 474–480. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Bowker GC, & Star SL (1999). Sorting things out: Classification and its consequences . MIT Press. [ Google Scholar ]
• Bray GA, Kim KK, Wilding JPH, & World Obesity Federation. (2017). Obesity: A chronic relapsing progressive disease process. A position statement of the World Obesity Federation . Obesity Reviews , 18 ( 7 ), 715–723. [ PubMed ] [ Google Scholar ]
• Brownell KD, Kersh R, Ludwig DS, Post RC, Puhl RM, Schwartz MB, & Willett WC (2010). Personal responsibility and obesity: A constructive approach to a controversial issue . Health Affairs , 29 ( 3 ), 379–387. [ PubMed ] [ Google Scholar ]
• Burkhauser RV, & Cawley J (2008). Beyond BMI: The value of more accurate measures of fatness and obesity in social science research . Journal of Health Economics , 27 ( 2 ), 519–529. [ PubMed ] [ Google Scholar ]
• Cahnman WJ (1968). The stigma of obesity . The Sociological Quarterly , 9 ( 3 ), 283–299. [ Google Scholar ]
• Centers for Disease Control & Prevention. (2020). COVID-19 Racial and Ethnic Health Disparities . https://www.cdc.gov/coronavirus/2019-ncov/community/health-equity/racial-ethnic-disparities/index.html .
• Centers for Disease Control & Prevention. (2021). People with Certain Medical Conditions . https://www.cdc.gov/coronavirus/2019-ncov/need-extra-precautions/people-with-medical-conditions.html .
• Charrow A, & Yerramilli D (2018). Obesity as disease: Metaphysical and ethical considerations . Ethics, Medicine and Public Health , 7 , 74–81. [ Google Scholar ]
• Ciciurkaite G, & Perry BL (2018). Body weight, perceived weight stigma and mental health among women at the intersection of race/ethnicity and socioeconomic status: Insights from the modified labelling approach . Sociology of Health & Illness , 40 ( 1 ), 18–37. [ PubMed ] [ Google Scholar ]
• Dew K (2012). The cult and science of public health: A sociological investigation . Berghahn Books. [ Google Scholar ]
• Epstein S (2008). Inclusion: The politics of difference in medical research . University of Chicago Press. [ Google Scholar ]
• Fletcher I (2014). Defining an epidemic: The body mass index in British and US obesity research 1960–2000 . Sociology of Health & Illness , 36 ( 3 ), 338–353. [ PubMed ] [ Google Scholar ]
• Flint SW (2020). Stigmatizing Media Portrayal of Obesity During the Coronavirus (COVID-19) Pandemic . Frontiers in Psychology , 11 . [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Fourcade M (2016). Ordinalization: Lewis A. Coser memorial award for theoretical agenda setting 2014 . Sociological Theory , 34 ( 3 ), 175–195. [ Google Scholar ]
• Fox NJ (2012). The Body . Polity. [ Google Scholar ]
• Frank R (2007). What to make of it? The (Re) emergence of a biological conceptualization of race in health disparities research . Social Science & Medicine , 64 ( 10 ), 1977–1983. [ PubMed ] [ Google Scholar ]
• Gillespie C (2012). The experience of risk as ‘measured vulnerability’: Health screening and lay uses of numerical risk . Sociology of Health & Illness , 34 ( 2 ), 194–207. [ PubMed ] [ Google Scholar ]
• Gillespie C (2015). The risk experience: The social effects of health screening and the emergence of a proto-illness . Sociology of Health & Illness , 37 ( 7 ), 973–987. [ PubMed ] [ Google Scholar ]
• Greenhalgh S (2015). Fat-talk nation: The human costs of America’s war on fat . Cornell University Press. [ Google Scholar ]
• Gutin I (2018). In BMI we trust: Reframing the body mass index as a measure of health . Social Theory & Health , 16 ( 3 ), 256–271. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Gutin I (2019). Essential(ist) medicine: Promoting social explanations for racial variation in biomedical research . Medical Humanities , 45 ( 3 ), 224–234. [ PubMed ] [ Google Scholar ]
• Hales CM, Carroll MD, Fryar CD, & Ogden CL (2020). Prevalence of obesity and severe obesity among adults: United States, 2017–2018 . NCHS Data Brief , 360 . [ PubMed ] [ Google Scholar ]
• Harris KM, & Schorpp KM (2018). Integrating biomarkers in social stratification and health research . Annual Review of Sociology , 44 , 361–386. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hatzenbuehler ML, Keyes KM, & Hasin DS (2009). Associations between perceived weight discrimination and the prevalence of psychiatric disorders in the general population . Obesity , 17 ( 11 ), 2033–2039. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hatzenbuehler ML, Phelan JC, & Link BG (2013). Stigma as a fundamental cause of population health inequalities . American Journal of Public Health , 103 ( 5 ), 813–821. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hill MA, Sowers JR, & Mantzoros CS (2021). Commentary: COVID-19 and obesity pandemics converge into a syndemic requiring urgent and multidisciplinary action . Metabolism , 114 . [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hooper MW, Nápoles AM, & Pérez-Stable EJ (2020). COVID-19 and racial/ethnic disparities . JAMA , 323 ( 24 ), 2466–2467. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hu FB (2008). Measurements of adiposity and body composition. In Hu F (Ed.), Obesity epidemiology (pp. 53–83). Oxford University Press. [ Google Scholar ]
• Jastreboff AM, Kotz CM, Kahan S, Kelly AS, & Heymsfield SB (2019). Obesity as a disease: The obesity society 2018 position statement . Obesity , 27 ( 1 ), 7–9. [ PubMed ] [ Google Scholar ]
• Jauho M (2019). Patients-in-waiting or chronically healthy individuals? People with elevated cholesterol talk about risk . Sociology of Health & Illness , 41 ( 5 ), 867–881. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Jones DS, & Oppenheimer GM (2017). If the Framingham Heart Study Did Not Invent the Risk Factor, Who Did? . Perspectives in Biology and Medicine , 60 ( 2 ), 131–150. [ PubMed ] [ Google Scholar ]
• Jutel A (2001). Does size really matter? Weight and values in public health . Perspectives in Biology and Medicine , 44 ( 2 ), 283–296. [ PubMed ] [ Google Scholar ]
• Jutel A (2006). The emergence of overweight as a disease entity: Measuring up normality . Social Science & Medicine , 63 ( 9 ), 2268–2276. [ PubMed ] [ Google Scholar ]
• Jutel A (2011). Classification, disease, and diagnosis . Perspectives in Biology and Medicine , 54 ( 2 ), 189–205. [ PubMed ] [ Google Scholar ]
• Jutel A (2014). Putting a name to it: Diagnosis in contemporary society . Johns Hopkins University Press. [ PubMed ] [ Google Scholar ]
• Jutel A (2019). Diagnosis: Truths and tales . University of Toronto Press. [ Google Scholar ]
• Jutel A, & Buetow S (2007). A picture of health?: Unmasking the role of appearance in health . Perspectives in Biology and Medicine , 50 ( 3 ), 421–434. [ PubMed ] [ Google Scholar ]
• Kelly MP, & Russo F (2018). Causal narratives in public health: The difference between mechanisms of aetiology and mechanisms of prevention in non-communicable diseases . Sociology of Health & Illness , 40 ( 1 ), 82–99. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kragelund C, & Omland T (2005). A farewell to body-mass index? . The Lancet , 366 ( 9497 ), 1589–1591. [ PubMed ] [ Google Scholar ]
• Krieger N (2020). Enough: COVID-19, Structural Racism, Police Brutality, Plutocracy, Climate Change-and Time for Health Justice, Democratic Governance, and an Equitable, Sustainable Future . American Journal of Public Health , e1–e4. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kwok S, Adam S, Ho JH, Iqbal Z, Turkington P, Razvi S, Le Roux CW, Soran H & Syed AA (2020). Obesity: A critical risk factor in the COVID-19 pandemic . Clinical Obesity , 10 ( 6 ). [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kyle TK, Dhurandhar EJ, & Allison DB (2016). Regarding obesity as a disease: Evolving policies and their implications . Endocrinology and Metabolism Clinics , 45 ( 3 ), 511–520. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Le Brocq S, Clare K, Bryant M, Roberts K, & Tahrani AA (2020). Obesity and COVID-19: A call for action from people living with obesity . The Lancet Diabetes & Endocrinology , 8 ( 8 ), 652–654. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Loucks EB, Juster RP, & Pruessner JC (2008). Neuroendocrine biomarkers, allostatic load, and the challenge of measurement: A commentary on Gersten . Social Science & Medicine , 66 ( 3 ), 525–530. [ Google Scholar ]
• Lupton D (2006). Sociology and risk. In Mythen Gand Walklate S (Eds.), Beyond the risk society: Critical reflections on risk and human security (pp. 11–24). Open University Press. [ Google Scholar ]
• Maddox GL, Back KW, & Liederman VR (1968). Overweight as social deviance and disability . Journal of Health and Social Behavior , 287–298. [ PubMed ] [ Google Scholar ]
• Mann T, Tomiyama AJ, Westling E, Lew AM, Samuels B, & Chatman J (2007). Medicare’s search for effective obesity treatments: diets are not the answer . American Psychologist , 62 ( 3 ), 220–233. [ PubMed ] [ Google Scholar ]
• McDade TW (2008). Challenges and opportunities for integrative health research in the context of culture: A commentary on Gersten . Social Science & Medicine , 66 ( 3 ), 520–524. [ Google Scholar ]
• Monaghan LF (2007). Body Mass Index, masculinities and moral worth: Men’s critical understandings of ‘appropriate’ weight-for-height . Sociology of Health & Illlness , 29 ( 4 ), 584–609. [ PubMed ] [ Google Scholar ]
• Monaghan LF (2020). Coronavirus (COVID-19), pandemic psychology and the fractured society: A sociological case for critique, foresight and action . Sociology of Health & Illness , 42 ( 8 ), 1982–1995. [ PubMed ] [ Google Scholar ]
• Morning A (2011). The nature of race: How scientists think and teach about human difference . University of California Press. [ Google Scholar ]
• Murray S (2009). Marked as ‘pathological’:’Fat’bodies as virtual confessors. In Wright J and Harwood V (Eds.), Biopolitics and the ‘obesity epidemic’: Governing bodies (pp. 78–90). Routledge. [ Google Scholar ]
• Nicholls SG (2013). Standards and classification: A perspective on the ‘obesity epidemic’ . Social Science & Medicine , 87 , 9–15. [ PubMed ] [ Google Scholar ]
• Nuttall FQ (2015). Body mass index: Obesity, BMI, and health: a critical review . Nutrition Today , 50 ( 3 ), 117–128. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Oliver JE (2006). Fat politics: The real story behind America’s obesity epidemic . Oxford University Press. [ Google Scholar ]
• Papadopoulos S, & Brennan L (2015). Correlates of weight stigma in adults with overweight and obesity: A systematic literature review . Obesity , 23 ( 9 ), 1743–1760. [ PubMed ] [ Google Scholar ]
• Phelan JC, & Link BG (2015). Is racism a fundamental cause of inequalities in health? . Annual Review of Sociology , 41 , 311–330. [ Google Scholar ]
• Phelan SM, Burgess DJ, Yeazel MW, Hellerstedt WL, Griffin JM, & van Ryn M (2015). Impact of weight bias and stigma on quality of care and outcomes for patients with obesity . Obesity Reviews , 16 ( 4 ), 319–326. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Popkin BM, Du S, Green WD, Beck MA, Algaith T, Herbst CH, Alsukait RF, Alluhidan M, Alazemi N, & Shekar M (2020). Individuals with obesity and COVID-19: A global perspective on the epidemiology and biological relationships . Obesity Reviews , 21 ( 11 ). [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Posner RA, & Philipson TJ (2003). The Long-Run Growth in Obesity as a Function of Technological Change . Perspectives in Biology and Medicine , 46 , S87–S107. [ PubMed ] [ Google Scholar ]
• Puhl RM, & Heuer CA (2009). The stigma of obesity: A review and update . Obesity , 17 ( 5 ), 941–964. [ PubMed ] [ Google Scholar ]
• Puhl RM, & Heuer CA (2010). Obesity stigma: Important considerations for public health . American Journal of Public Health , 100 ( 6 ), 1019–1028. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Puhl RM, Lessard LM, Larson N, Eisenberg ME, & Neumark-Stzainer D (2020). Weight stigma as a predictor of distress and maladaptive eating behaviors during COVID-19: Longitudinal findings from the EAT study . Annals of Behavioral Medicine , 54 ( 10 ), 738–746. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Puhl RM, Moss-Racusin CA, Schwartz MB, & Brownell KD (2008). Weight stigmatization and bias reduction: Perspectives of overweight and obese adults . Health education research , 23 ( 2 ), 347–358. [ PubMed ] [ Google Scholar ]
• Puhl R, & Suh Y (2015). Health consequences of weight stigma: Implications for obesity prevention and treatment . Current Obesity Reports , 4 ( 2 ), 182–190. [ PubMed ] [ Google Scholar ]
• Rosenberg CE (2002). The tyranny of diagnosis: Specific entities and individual experience . The Milbank Quarterly , 80 ( 2 ), 237–260. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Rothman KJ (2008). BMI-related errors in the measurement of obesity . International Journal of Obesity , 32 ( 3 ), S56–S59. [ PubMed ] [ Google Scholar ]
• Saguy AC (2012). What’s wrong with fat? . Oxford University Press. [ Google Scholar ]
• Scambler G (2009). Health-related stigma . Sociology of Health & Illness , 31 ( 3 ), 441–455. [ PubMed ] [ Google Scholar ]
• Schafer MH, & Ferraro KF (2011). The stigma of obesity: Does perceived weight discrimination affect identity and physical health? . Social Psychology Quarterly , 74 ( 1 ), 76–97. [ Google Scholar ]
• Schmidt H, Gostin LO, & Williams MA (2020). Is it lawful and ethical to prioritize racial minorities for COVID-19 vaccines? . JAMA , 324 ( 20 ), 2023–2024. [ PubMed ] [ Google Scholar ]
• Sharma AM, & Campbell-Scherer DL (2017). Redefining obesity: Beyond the numbers . Obesity , 25 ( 4 ), 660–661. [ PubMed ] [ Google Scholar ]
• Shim JK (2002). Understanding the routinised inclusion of race, socioeconomic status and sex in epidemiology: The utility of concepts from technoscience studies . Sociology of health & illness , 24 ( 2 ), 129–150. [ Google Scholar ]
• Shugart HA (2016). Heavy: The obesity crisis in cultural context . Oxford University Press. [ Google Scholar ]
• Smart A, & Weiner K (2018). Racialised prescribing: Enacting race/ethnicity in clinical practice guidelines and in accounts of clinical practice . Sociology of Health & Illness , 40 ( 5 ), 843–858. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Stokes A, & Preston SH (2016). How dangerous is obesity? Issues in measurement and interpretation . Population and Development Review , 42 ( 4 ), 595–614. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Sutin AR, Stephan Y, & Terracciano A (2015). Weight discrimination and risk of mortality . Psychological Science , 26 ( 11 ), 1803–1811. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Timmermans S, & Berg M (2003). The gold standard: The challenge of evidence-based medicine . Temple University Press. [ Google Scholar ]
• Timmermans S, & Haas S (2008). Towards a sociology of disease . Sociology of Health & Illness , 30 ( 5 ), 659–676. [ PubMed ] [ Google Scholar ]
• Tomiyama AJ, Carr D, Granberg EM, Major B, Robinson E, Sutin AR, & Brewis A (2018). How and why weight stigma drives the obesity ‘epidemic’ and harms health . BMC medicine , 16 ( 1 ), 1–6. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Townsend MJ, Kyle TK, & Stanford FC (2020). Commentary: COVID-19 and obesity: exploring biologic vulnerabilities, structural disparities, and weight stigma . Metabolism-Clinical and Experimental , 110 . [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Tylka TL, Annunziato RA, Burgard D, Daníelsdóttir S, Shuman E, Davis C, & Calogero RM (2014). The weight-inclusive versus weight-normative approach to health: Evaluating the evidence for prioritizing well-being over weight loss . Journal of Obesity . [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Wu K (2020). Studies Begin to Untangle Obesity’s Role in Covid-19 . The New York Times. https://www.nytimes.com/2020/09/29/health/covid-obesity.html . [ Google Scholar ]
• Yancy CW (2020). COVID-19 and african americans . JAMA , 323 ( 19 ), 1891–1892. [ PubMed ] [ Google Scholar ]
• Yates-Doerr E (2013). The mismeasure of obesity. In McCullough MB and Hardin JA (Eds.), Reconstructing obesity research: The measure of meaning, the meaning of measures (pp. 49–70). Berghahn Books. [ Google Scholar ]

Official websites use .gov

A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS

A lock ( ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Case Study: Plotting and interpreting BMI-for-age using the English system

A case study of ‘charlene’.

Step 1. Obtain accurate weights and stature measurements

Charlene, a girl, comes in for a visit.

Here is Charlene’s basic information:

Date of Birth (DOB): 10/30/96 Date of Visit (DOV): 11/5/00 Weight = 40 pounds 4 ounces Height = 40 3/4 inches

Step 2. Select the appropriate growth chart

Based upon the above information, select a growth chart:

Please choose one Boys 2 to 20 BMI-for-age Girls 2 to 20 BMI-for-age Boys 2 to 20 weight-for-age Girls 2 to 20 weight-for-age Boys 2 to 20 stature-for-age Girls 2 to 20 stature-for-age That’s incorrect. Please try again.

That’s correct!

Because Charlene is a normal 4-year old, a standing height was obtained. The appropriate chart to use is the ‘Girls 2 to 20 BMI-for-age’ chart.

Step 3. Record the data

The data entry table on the clinical growth chart can be completed with information relevant to the growth chart. Enter the missing data on the Data Entry Table and then click the ‘Submit’ button.

Charlene Date of visit: 11/05/00 Child’s age: 4 Weight: 40 lbs 4 oz Height: 40 3/4 inches

Step 4. Calculate BMI

Now calculate BMI at the time of Charlene’s clinic visit.

Date of visit: 11/05/00 Child’s age: 4 Weight: 40 lbs 4 oz Height: 40 3/4 inches

Convert ounces and fractions to decimals: Weight of 40 lbs and 4 oz = 40.25 lbs (16 ounces = 1 pound, so 4 oz / 16 oz = 0.25)

Height = 40.75 inches

BMI = (weight / height / height) x 703

BMI = (40.25 lbs/40.75 in/40.75 in) x 703 = 17.0

Note: There is a difference of 0.1 between the BMI calculations when using the metric system (17.1) versus the English system (17.0). This is due to the conversion factor of 703.

Enter the BMI on the Data Entry Table and click ‘Submit’:

That’s incorrect. Please try again.

BMI is incorrect. Please try again.

That’s right.

Step 5. Plot measurements

All of the necessary information is recorded and Charlene’s BMI can be plotted. On the BMI-for-age chart, find Charlene’s age on the horizontal axis and visually draw a vertical line up from that point. Then find her BMI on the vertical axis and visually draw a horizontal line across from that point. The point where the two intersect represents Charlene’s BMI-for-age.

Click on the correctly plotted BMI-for-age chart from the three options below:

That’s the correct chart!

Step 6. Interpret the plotted measurements

Since Charlene’s BMI-for-age falls between the 85th and 90th percentile curves, this means that of 100 children with the same gender and age as Charlene, 85 to 90 children will have a BMI-for-age lower than Charlene’s and 10 to 15 children will have a BMI-for-age greater than hers. Because Charlene’s BMI-for-age is above the 85th percentile, and lower than the 95th percentile, she is…

Please choose one underweight normal overweight obese That’s incorrect. Please try again.

Since Charlene’s BMI-for-age falls between the 85th and 90th percentile curves, she is overweight.

Congratulations! You have successfully calculated, plotted and interpreted Charlene’s BMI-for-age.

Thank you for completing the module. Click here to visit the references and resources page, or click below to return to the Main Modules page.

Exit Notification / Disclaimer Policy

• The Centers for Disease Control and Prevention (CDC) cannot attest to the accuracy of a non-federal website.
• Linking to a non-federal website does not constitute an endorsement by CDC or any of its employees of the sponsors or the information and products presented on the website.
• You will be subject to the destination website's privacy policy when you follow the link.
• CDC is not responsible for Section 508 compliance (accessibility) on other federal or private website.

• Case Study – Margaret

Margaret is 72 years old and has COPD. She knows that she needs to allow herself enough time to get ready whenever she goes out. On this particular morning she has an early appointment at the hospital that she has been waiting several months for. She is getting picked up by her son John and she has not slept well for fear of sleeping in. She feels breathless when she wakes up and it takes her longer to take her medication and get dressed even although she had everything laid out the night before.

Her son, John arrives before she is ready and says that she will need to watch her time or she will be late because the traffic is bad. Margaret starts to worry that she will not make it in time, she tries to quicken her pace but her breathlessness increases, her heart races and she does not feel well.   

John tries to speak to her but she struggles to get her words out and starts to panic that she is not going to get a breath. Her heart is racing so fast she is scared that she might have a heart attack.   

John doesn’t know what to do, normally his sister takes his mum to appointments, he also starts to panic that his mum is seriously ill. He calls an ambulance.

What is happening to Margaret?

What is margaret doing.

Focusing more and more on breathing. Not using breathing techniques. Calling for help.

What is Margaret thinking?

“I can’t breathe” “John is panicking, it must be serious” “I am going to miss my appointment and my doctor will be annoyed” “I am going to die”

How is Margaret feeling?

Worried. Scared. Increasing to panic...

What is happening to Margaret physically?

Breathing faster and feeling More breathless. Heart pounding. Dizzy. Feeling sick.

Margaret is seen by a doctor at the hospital. She tells her about the episode leading to her admission and also her worry. Whilst assessing Margaret the doctor notices that she frequently yawns and sighs and appeared to hold her breath a lot. She notices that Margaret’s mood seems a bit flat and she is quite weepy throughout their meeting. She asks Margaret to complete a short  mental health screening measure  and on the basis of this asks Margaret if she would like to attend the Clinical Psychologist for further assessment of her mood.  She may be depressed and anxious – both are treatable. She also refers Margaret to a respiratory physiotherapist for breathing retraining. 

Margaret sees the physiotherapist... 

Margaret tells the physiotherapist that she has been a bit lonely since her husband passed away and that although she enjoys going out for coffee with her friend it can be a struggle for her at times due to her breathing. Her son John and daughter Kate live near by but she doesn’t like to bother them as they are both working. The physiotherapist helps Margaret understand how her feelings are linked to poor breathing habits and symptoms. He explains that her bad breathing habits at home such as frequent sighing and yawning means that the total volume of air moving in and out of her airways is excessive. This  results in chronic low levels of carbon dioxide and the other symptoms reported by Margaret. He advises Margaret to avoid sighing and yawning and to practice relaxed slow, shallow breathing when feeling sad or anxious. 

Margaret sees the clinical psychologist...

The clinical psychologist uses a cognitive behavioral approach to help Margaret identify her anxious thoughts and triggers for panic. They spend time learning relaxation techniques in addition to the breathing techniques she has been shown by the physiotherapist. Margaret becomes more aware of her body symptoms and sets herself small goals to test out what she can do. Margaret also agrees to be referred to a local support group. Over time Margaret’s mood improves. She enjoys the regular contact with the group. She finds that her frightening symptoms can be avoided or controlled by distraction and keeping occupied as well as practicing her new anxiety management and breathing techniques.

Margaret’s outcome

Diaphragmatic breathing. Stopping & resting. Distraction techniques. Relaxation & changing posture.

“I am safe”. “My breathing will calm down”. “Nothing awful is going to happen”.

Calmer. Less irratable. More in control.

Breathing is slowing down,. heart rate lower, less tense

• Topic 1 - Anatomy and Physiology
• Topic 2 - Assessment and common lung diseases
• Oxygen therapy
• Pulmonary Rehabilitation (PR)
• Breathing techniques
• Cough control
• Emotional or mental wellbeing
• Mental wellbeing and respiratory conditions
• Recognising mental health problems
• Communication
• Low mood and depression
• Management of depressive symptoms
• The panic & breathlessness spiral
• Recognising physical signs of anxiety
• Physiological effects of fight/flight response
• Fight or flight
• Cognitive Behavioural Therapy
• How can you reduce anxiety?
• Anxiety and breathlessness
• What is a panic attack?
• Theories about panic attacks and breathing
• Mental health self-help
• Topic 4 - Management

Copyright © 2018-2024 Chest Heart & Stroke Scotland and The University of Edinburgh . Some rights reserved. Site published by the Interactive Content Team , Information Services, The University of Edinburgh